Search

In this paper, a new approach for obstacle vehicle path prediction, which is important for advanced driver assistance systems (ADAS) and autonomous vehicles, is proposed based on a deep neural network. In order to analyze sequential sensor data, a recurrent neural network (RNN) is used and the input data for RNN is drawn from three sensors: LIDAR, camera and GPS. These sensor data are obtained experimentally with real vehicles. In addition, deep ensemble is used for robustness of the estimation and acquisition of the uncertainty. The predicted path of the proposed method is continuous and it predicts both short-term and long-term path with a single algorithm. The size of the network model is small, but it shows good performance in predicting future trajectory of obstacle vehicles.

Even though some of the driver assistant systems have been commercialized to provide safety and convenience to the driver, they can be applied for autonomous driving in limited situations such as highways. In this paper, we propose a supervisor agent that can enhance the driver assistant systems by using deep distributional reinforcement learning. The supervisor agent is trained using end-to-end approach that directly maps both a camera image and LIDAR data into action plan. Because the well-trained network of deep reinforcement learning can lead to unexpected actions, collision avoidance function is added to prevent dangerous situations. In addition, the highway driving case is a stochastic environment with inherent randomness and, thus, its training is performed through the distributional reinforcement learning algorithm, which is specialized for stochastic environment. The optimal action for autonomous driving is selected through the return value distribution. Finally, the proposed algorithm is verified through a highway driving simulator, which is implemented by the Unity ML-agents.

Adaptive cruise control (ACC) considering vehicle performances becomes a core technology for autonomous vehicle to adjust the vehicle speed while maintaining a safe distance from preceding vehicles...

In this paper, an active front steering (AFS) controller with a new variable steering ratio (VSR) map is proposed to enhance the vehicle stability and consider the driver’s steering comfort. The properties of the steering sensitivity and the yaw rate gain are investigated and utilized to construct a 3D VSR map in terms of the steering wheel angle and the longitudinal vehicle velocity. Based on this 3D VSR map, the AFS controller is designed by adopting the model predictive control (MPC) approach. An optimum solution of the MPC controller is determined by taking into account the AFS actuator limits. The proposed algorithm is validated in simulations by using the commercial software, CarSim and MATLAB/Simulink.

The Autonomous Emergency Braking (AEB) systems have been actively studied for the safety enhancement and commercialized for the past few years. Because the driver tends to overly rely upon active safety systems, AEB needs to be designed to reflect the real road situations such as various road slope and friction coefficient. In this study, an AEB control algorithm is proposed to compensate for the effects of the slope and the friction of road. Based on the maximum possible deceleration for the real road conditions, the minimum braking distance is described with margin parameters for AEB activation control. The deceleration controller with a feedforward term is designed to avoid the collision during AEB operation on real road conditions. The proposed algorithm is validated in simulations first and the experimental verification is performed in the various slope conditions.

For the path planning of autonomous vehicles, it is important to predict the future trajectory of the surrounding vehicles. However, predicting future trajectory is difficult because it needs to consider the invisible interaction between the vehicles in a dynamic driving environment. In this paper, a new approach, which considers the interaction between surrounding vehicles, is proposed for accurate prediction of the future trajectory. The proposed method provides continuous predicted trajectories over time in the longitudinal and lateral directions, respectively. The deep ensemble technique is also used to predict the uncertainty of the estimated trajectory. This paper performs the training and verification of the algorithm using NGSIM dataset, which is the vehicle driving data obtained through actual vehicle driving.

This paper presents online-capable deep learning model for probabilistic vehicle trajectory prediction. We propose a simple encoder-decoder architecture based on multi-head attention. The proposed model generates the distribution of the predicted trajectories for multiple vehicles in parallel. Our approach to model the interactions can learn to attend to a few influential vehicles in an unsupervised manner, which can improve the interpretability of the network. The experiments using naturalistic trajectories at highway show the clear improvement in terms of positional error on both longitudinal and lateral direction., Comment: 6 pages, 5 figures, 2020 IEEE Intelligent Vehicles Symposium (IV)

For autonomous vehicles, lane changes on crowded roads are difficult to be performed without interactions and cooperation between vehicles. This paper proposes a novel method to learn interaction and cooperate between the multiple vehicles to solve the complex traffic problem through Multi-Agent Reinforcement Learning (MARL). The proposed network is designed based on the interaction network to learn optimal control strategies considering interaction between vehicles. By applying the proposed algorithm, the network can control and train the agents regardless of the number of agents. It is a practical advantage because the number of the vehicles is constantly changed in the real environment. The proposed method is evaluated in the connected car environment where all vehicles can exchange information with each other.

Interaction-aware prediction is a critical component for realistic path planning that prevents automated vehicles from overly cautious driving. It requires to consider internal states of other driver such as driving style and intention, which the automated vehicle cannot directly measure. This paper proposes a probabilistic driver model for response prediction given the planned future actions of automated vehicle. The drivers internal states are considered in an unsupervised manner. The prediction model utilizes mixture density network to estimate future acceleration and yaw-rate profile of interacting vehicles. The proposed method is evaluated by using real-world trajectory data.

A planar motion control system is proposed for autonomous vehicles not only to follow the lanes, but also to avoid collisions by braking, accelerating, and steering. The supervisor is designed first to determine the desired speed and the risk of the maneuvering due to road boundaries and obstacles. In order to allow lane changes on multi-lane roads, the model predictive controller is formulated based on the probabilistic non-convex optimization. The micro-genetic algorithm is applied to calculate the target speed and target steering angle in real time. A software-in-the-loop unit is constructed with the Rapid Control Prototyping device in the vehicle communication environment. The performance of the proposed system is verified for various collision avoidance scenarios and the simulation results demonstrate the safe and effective driving performance of autonomous vehicles with no collision on multi-lane road.

Estimation algorithms for road slope angle and vehicle mass are presented for commercial vehicles. It is well known that vehicle weight and road grade significantly affect the longitudinal motion of a commercial vehicle. However, it is very difficult to measure the weight and road slope angle in real time because of lack of sensor technology. In addition, the total weight of a commercial vehicles varies depending on the freight. In this study, the road grade and vehicle mass estimation algorithms are proposed using the RLS (Recursive Least Square) method and only the in-vehicle sensors. The proposed algorithms are verified in experiments using a commercial vehicle under various conditions.

This paper proposes an active roll control system for passenger cars. The roll actuator with electric motor is expected to replace the hydraulic roll control actuator with the better performance. In order to meet those expectations, highly precise control methods are needed for adjusting the roll motion of the vehicle. In this study, roll dynamics of vehicle and the properties of active roll actuators are investigated first including the latency of the actuator. The identification method is designed to estimate the key parameters of the suspension system. The reference model method is proposed to determine the target roll states and the model predictive control (MPC) method is adopted to control the rolling motion. The roll moment distribution method is also designed between the front and rear actuators considering the load transfer. The proposed control system is validated via simulations and experiments.

This research was supported by Korea Evaluation Institute of Industrial Technology(KEIT) grant funded by the Korea government(MOTIE) (No. 10052375, Test Scenario Study for Supporting Autonomous Driving Technology Development).

Recent research suggests the various applications of Model Predictive Control on vehicle systems. In numerous cases, nonlinear tire models such as the Magic Formula, which are highly complex and are more detailed than necessary, are used. This paper presents a nonlinear tire model that excludes the region of negative slope but expresses the nonlinear properties of tire well enough for tracking the lane of a racing course. The proposed inverse tire model can also be used to calculate the slip angle from the tire force. Thus, the model can be utilized to design the Model Predictive Controller.

Recently, an active safety system based on vehicle-to-vehicle communication was introduced to minimize the threat of car accidents on the road and to overcome the limitations of the current sensor-based advanced driver assist system. In order to implement the system based on vehicle-to-vehicle communication, target classification is the key layer to be developed. In this study, based on the transmitted path history of the remote vehicles, the road geometry around the host vehicle is reconstructed without in-vehicle sensors such as vision cameras. A tracking algorithm for the remote vehicle is formulated in order to predict its position continuously by using an extended Kalman filter. A local map is obtained with an outlier filter, and the target classification algorithm is designed from vehicle-to-vehicle communication data. The proposed algorithms are validated by simulations and experiments carried out using test vehicles.

This paper presents a fault-tolerant brake torque controller for four-wheel-distributed braking systems with in-wheel motors and Electro-Mechanical Brakes (EMB). Mechanical and electrical faults can degrade the performance of the EMB actuators and, thus, their effects need to be compensated in vehicle dynamics level. In this study, the faults are identified as performance degradation and expressed by the gains of each actuator. Assuming the brake force distribution and the regenerative braking ratios, the over-actuated braking system is simplified into a two-input system. A sliding mode controller is designed to track the driver’s braking and steering commands, even if there exist faults in EMBs. In addition, adaptive schemes are constructed to achieve the fault-tolerant control in braking. The proposed controller and strategies are verified in the EMB HILS (Hardware-in-loop-simulation) unit for various conditions.

It is well known that laser scanner has better accuracy than stereo vision in detecting the distance and velocity of the obstacles, whereas stereo vision can distinguish the objects better than the laser scanner. These advantages of each sensor can be maximized by sensor fusion approach so that the obstacles in front can be detected accurately. In this paper, high-level sensor fusion for the laser scanner and stereo vision is developed for object matching between the sensors. Time synchronization, object age, and reordering algorithms are designed for robust tracking of the objects. A time-delay update algorithm is also developed to determine the process time delay of the laser scanner. The expanded laser scanner data at every 1 ms is predicted by Kalman filter and is matched with the stereo vision data at every 66 ms. A cost function is formulated to describe the object matching similarity between the sensors, and the best matching candidate is selected for the minimum cost function. The proposed matching algorithms are verified experimentally in field tests of various maneuvering cases.

This paper presents a novel sideslip angle estimation scheme combining deep neural network and nonlinear Kalman filters. The deep neural network contains a recurrent neural network with long short-term memory which is effective for analyzing sequential sensor data and deep ensemble which is used for robustness of the estimation and acquisition of the uncertainty of the estimate. The deep neural network is trained using input sets which consist of on-board sensor measurements (yawrate, velocity, steering wheel angle and lateral acceleration) and provides sideslip angle estimate and its uncertainty. The estimate of deep neural network is used as a new measure in the nonlinear Kalman filters and its uncertainty is used to make an adaptive measurement covariance matrix. The algorithm is verified through both simulation and experiment. The performance with the proposed method is analyzed in terms of the root mean squared error (RMSE) and maximum error (ME) as compared to the case where nonlinear Kalman filter or deep neural network is utilized individually. The results demonstrate the effectiveness of the proposed solution.

In this paper, we propose a novel autonomous highway driving framework using optimal trajectory generation with an adaptive potential field model. While the existing potential field models for trajectory generation are hard to consider dynamic aspects of obstacles, the proposed potential field model overcomes such limitation by changing the risk size of the potential field. To express the risk function of moving object, the constant time gap policy is utilized. By combining the adaptive potential field and optimal trajectory generation scheme, the proposed driving framework successfully makes autonomous vehicles perform a variety of highway driving functionalities such as lane keeping, lane changing, distance keeping and collision avoidance in critical situations.

Estimating the maximum road friction coefficient with high reliability in various driving situation is one of the most significant issue in the field of automotive research. Numerous study has been done in this field, however, because of the several limitations and problems, researches in this field are still active. This paper uses a deep learning method to estimate the maximum road friction coefficient. The network of this study is mainly composed of convolutional neural network, recurrent neural network with deep ensemble architecture. In addition, through Prioritized Batch Selection (PBS), which is proposed in this paper, the training result is dramatically enhanced. The performance of the proposed estimator is verified in simulation of test driving scenarios.

With the commercialization of various Driver Assistance Systems (DAS), those vehicles have some autonomous functions like Smart Cruise Control (SCC) and Lane Keeping System (LKS). It is believed that autonomous driving can be achieved by combining the DAS functions in the limited situations such as on highways. However, in order to coordinate the DAS functions for autonomous driving, a supervisor is needed to select an appropriate DAS function. In this paper, we propose a method for training a supervisor that selects proper DAS by deep reinforcement learning. The driving policy operates based on camera images and LIDAR data that are accessible in autonomous vehicles. Therefore, deep reinforcement learning network model is designed to analyze both camera image and LIDAR data. This system aims to drive in simulated traffic situation of highway without collision and with high speed. Unlike the systems which learn how to throttle, brake and steering directly, the proposed method can guarantee safe driving because the learned driving policy is based on the existing commercialized DAS functions. In order to verify the algorithms, a simulation tool is developed using Unity for highway environment with multiple vehicles and autonomous driving performance is compared with the proposed supervisor.

In the past, the development of an adaptive cruise control (ACC) algorithm considering fuel efficiency and the development of an ACC system considering performances such as fuel efficiency, ride comfort and trackability have been carried out. In addition, research on vehicle and ACC parameters optimization considering fuel efficiency, ride comfort, trackability and safe distance have been carried out. However, in real world, vehicle sprung mass and center of gravity are changed due to the number of vehicle occupants, and there are uncertainties in vehicle parameters such as tire radius, tire spring constant and so on. Therefore, ACC should be designed considering uncertainties due to the variations of vehicle parameters.

Semi-trailer trucks generate the articulation angle during turning situation due to structural characteristics. This angle can cause the trailer to deviate from the road boundary and, thus, lead to collision with structures and/or vehicles on the roadway. In this study, a narrow passage assistant system is proposed using rear steering so that the semi-trailer truck is controlled inside the lane boundary even in tight turning circumstances. Based on the measured lane information using camera sensors, the model predictive control is applied for predicting the future state and tracking the reference path. The distribution rule for steering angles of the multi-axle trailer is described such that each wheel of the trailer has the same turning center and their tire slip angles are consistent. The proposed system is verified using commercial software and its performance demonstrates the effectiveness of rear steering in semi-trailer truck.

The slope angle is one of the most influential factors which limit the control performance and the energy efficiency of not only internal-combustion engine vehicles but also electrified vehicles such as hybrid vehicles and plug-in hybrid electric vehicles. However, measuring the slope angle is not easy because slope sensors are not sufficiently accurate and they are too expensive to implement in vehicles. In this study, the climbing angle as well as the slope angle are considered in the presence of yawing motion on a sloping road, and both angles are estimated. In order to describe the combined vehicle motions when turning on a sloping road, a new three-dimensional dynamics model is proposed including the heading angle with respect to the road slope orientation. The extended Kalman filter is designed on the basis of the three-dimensional dynamics model on a sloping road, and the key parameters for the sloping road are updated using the geometry relationship. The proposed estimation method is evaluated using simulations and field tests with a hybrid electric vehicle.

In this paper, a regenerative electromagnetic brake, which has the same structure as conventional electromagnetic brakes and the same principle as self-excited induction generators, is introduced, and a miniature model is designed. The most important part of the design stage is to choose proper equivalent circuit parameters for self-excitation within a given operating speed range. This paper describes a 3-D map, called a CN-map, that consists of the rotating speed, capacitance, and stator resistance to verify the possibility of self-excitation during the design stage. A miniature model is designed using the map, and the experimental results obtained with the miniature model confirm the feasibility of the design process.

Deep learning is a fast-growing field of research, in particular, for autonomous application. In this study, a deep learning network based on various sensor data is proposed for identifying the roads where the vehicle is driving. Long-Short Term Memory (LSTM) unit and ensemble learning are utilized for network design and a feature selection technique is applied such that unnecessary sensor data could be excluded without a loss of performance. Real vehicle experiments were carried out for the learning and verification of the proposed deep learning structure. The classification performance was verified through four different test roads. The proposed network shows the classification accuracy of 94.6% in the test data.

In controlling the longitudinal motion of electrified vehicles such as hybrid vehicles and PHEV (Plug-in Hybrid Electric Vehicles), the variation of the driving resistance loads (or driving loads) such as road grade and actual vehicle mass, is the most influential factor which limits the control performance. Measuring the driving load is not impossible, but it is costly since additional sensors have to be mounted on the vehicle. In this study, methods for estimating vehicle mass and road grade are designed to compensate for the driving loads. The proposed methods are verified using simulation tools and then evaluated experimentally.

This paper describes an autonomous vehicle A1 that won the Hyundai Motor Group 2012 Autonomous Vehicle Competition. The A1 was developed for autonomous on-road and off-road driving conditions without driver intervention. The autonomous driving system consists of four parts that are localization, perception, planning, and control. Localization which estimates the ego-vehicle position on the map should be first performed to autonomously drive the A1. A perception algorithm detects and recognizes the objects around the ego-vehicle are also important to prevent collisions with obstacles and road departure. A planning algorithm generates the drivable motion of the A1 based upon previous information from the localization and perception system. Subsequently, a vehicle control algorithm calculates the desired steering, acceleration and braking control commands based on the information from the planning algorithm. This paper also presents the entire system architecture that the A1 used to accomplish all the required missions in the 2012 Autonomous Vehicle Competition.

This paper presents the lane change system for collision avoidance. The proposed algorithm for the collision avoidance consists of path generation and path following. Using a calculated TTC (Time to Collision), partial braking is operated and collision avoidance path is generated considering relative distance, velocity and acceleration. Based on the collision avoidance path, desired yaw angle and yaw rate are calculated for the automated path following. The lateral controller is designed by a Lyapunov function approach using 3 D.O.F vehicle model and vehicle parameters. The required steering angle is determined from wheel velocity, longitudinal and lateral velocity in order to follow the desired yaw angle and yaw rate. This system is developed MATLAB/Simulink and its performance is evaluated using the commercial software CarSim.

Brake systems of the future, including BBW (Brake-by-Wire), are in development in various forms. In one of the proposed hydraulic BBW systems, an electric booster system replaces the pneumatic brake booster with an electric motor and a rotational-to-linear motion mechanism. This system is able to provide improved braking performance by the design of controllers with precise target pressure tracking and control robustness for better system reliability. First, a sliding mode controller is designed using the Lyapunov function approach to secure the robustness of the system against both the model uncertainty and the disturbance caused by the master cylinder and mechanical components. Next, a simulation tool is constructed to validate the electric booster system with the proposed controller. Finally, the electric booster system is implemented into an actual brake ECU and installed in a vehicle for testing under various braking conditions. The experimental results demonstrate that the proposed controller produces faster pressure build-up performance than the conventional brake system, and its tracking performance is sufficient to ensure comfortable braking.

In the future, the conventional hydraulic brake system in automobiles will be removed and replaced by an electrically operated brake system called brake-by-wire. The brake-by-wire units, such as the EMB (Electro-Mechanical Brake), provide better braking performance by directly controlling the brake motor and are environmentally friendly because they do not use hydraulic fluid. For implementation of the EMB systems, reliable and robust fault detection and diagnosis methods become increasingly important. In this study, a sensor fault diagnosis method is proposed with parity space and observer approaches to detect faults in the motor current sensor, speed (or position) sensor and clamping force sensor. The proposed method is verified through a closed-loop simulation using Matlab/Simulink, and the simulation result is compared with the HILS bench test results.

최근 소형 트럭이나 SUV(Sport Utility Vehicle) 차 량이 증가함에 따라 전복으로 인한 차량 사고를 줄이기 위하여, 최근 ESC(Electronic Stability Control)와 같은 ARP(Active Rollover Prevention) 제 어기가 상용화되고 있다. 이러한 제어기의 제어 성능은 차량의 관성 파라미터, 특히 차량의 질량 과 무게 중심의 높이에 민감하다. 이러한 차량 의 관성 파라미터는 승객의 인원수나 탑재된 하중 의 무게나 위치에 따라 달라진다. 특히, SUV 차량 은 지붕에 자전거나 보트 등의 무거운 짐을 탑재 하는 경우가 많기 때문에 제어기를 설계할 때, 이 러한 각각의 상황에 따라 변화하는 제어기가 고려 되어야 한다. 그러나 OEM 에서 제어기를 최악의 상황에 맞추어 안전도를 고려한 내용으로 설계하 고 있으며, 이런 경우 운전자의 의지가 급격한 또 는 불안정한 운동을 야기하는 경우에 제어기 작동 시 안정성은 확보된다고 할 수 있다. 하지만 때로 는 차량이 안정함에도 불필요한 개입을 통해 핸들 링감을 저하하는 원인으로 볼 수 있으며 설계 기 법은 종종 ESC 제어기에서의 차량의 핸들링 성능 을 현저하게 떨어뜨리는 원인이 되고 있다. 이러 한 문제점을 개선하기 위한 하나의 대책은 ARP 제어기에서 차량의 관성 파라미터에 대한 정확한 정보를 이용하는 것이다.

Collision detection plays a key role in collision mitigation system. The malfunction of the collision mitigation system can result in another dangerous situation or unexpected feeling to driver and passenger. To prevent this situation, the collision time, offset, and collision decision should be determined from the appropriate collision detection algorithm. This study focuses on a method to determine the time to collision (TTC) and frontal offset (FO) between the ego vehicle and the target object. The path prediction method using the ego vehicle information is proposed to improve the accuracy of TTC and FO. The path prediction method utilizes the ego vehicle motion data for better prediction performance. The proposed algorithm is developed based on laser scanner. The performance of the proposed detection algorithm is validated in simulations and experiments.

The electronic wedge brake is one of the brake-by-wire systems with a self-energizing effect. It is attractive because it can produce enough braking torque with the 12-voltage system. However, the electronic wedge brake cannot be implemented unless the self-energizing effect is effectively controlled. In this study, the electronic wedge brake is modeled into dynamic equations, and a sliding mode controller is designed based on the model. The clamping force is estimated based on the simplified electronic wedge brake model and the contact point detection algorithm is also provided. The performance of the proposed controller is verified in simulations and experiments using a prototype electronic wedge brake.

Collision detection plays a key role in collision mitigation systems. The malfunction of a collision mitigation system can result in a dangerous or unexpected situation for the driver and passengers. In order to prevent this situation, reliable collision detection algorithms are essential in terms of the collision time, the overlap and the collision decision. This study focuses on the reliable determination of the time to collision and the frontal overlap between the ego vehicle and the objects. The path prediction method using information on the ego vehicle is proposed to improve the accuracies of the time to collision and the frontal overlap calculations. The proposed algorithm is developed on the basis of a multi-layer laser scanner. The procedure of collision detection includes three steps. The first step is to select a high-risk object among multiple objects, the second step is to judge the time to collision and the frontal overlap to the high-risk object and, finally, the third step is to decide whether a crash is imminent or not. The performance of the proposed detection algorithm is validated in simulations and experiments.

EMB (Electro-Mechanical Brake) systems can provide improved braking and stability functions such as ABS, EBD, TCS, ESC, BA, ACC, etc. For the implementation of the EMB systems, reliable and robust fault detection algorithm is required. In this study, a model-based fault detection algorithm is designed based on the analytical redundancy method in order to monitor current and force sensor faults in EMB systems. A state-space model for the EMB is derived including faulty signals. The fault diagnosis algorithm is constructed using the analytical redundancy method. Observer is designed for the EMB and the fault detectability condition is examined based on the residual analysis. The performance of the proposed model-based fault detection algorithm is verified in simulations. The effectiveness of the proposed algorithm is demonstrated in various faulty cases.

For a class of uncertain nonlinear systems with sector-bounded nonlinearity, a proportional-integral state observer is designed in a dissipativity framework. The dissipativity ensures the bounded estimation errors against the bounded disturbance and an integral term in the observer structure can provide additional degrees of freedom in the observer design that can be used to improve the estimation performance. In this study, a dissipative condition for the proposed observer is found in terms of a linear matrix inequality (LMI). In order to specify the dissipativity, the L2 gain between the disturbance and the weighted sum of estimation errors is adopted as the supply rate in the dissipativity framework. In addition, the notion of the proposed design method is extended to the exponential dissipativity. The effectiveness of the proposed observer is demonstrated through a numerical example.

Monitoring of lateral vehicle motion is very useful in many active safety applications such as yaw stability control and rollover prevention. Lateral velocity and sideslip angle are regarded as the most important motion variables. However, it is not feasible during vehicle operation to directly measure them due to the high cost of sensors, limitations to sensor technology, etc. Therefore, the monitoring of lateral vehicle motion can be realized by using either accurate monitoring models or improved estimation algorithms. In this paper, monitoring models are obtained such that they utilize not only the vehicle dynamics but the cornering stiffness and the roll motion as well. Based on these monitoring models, three monitoring systems are constructed to estimate the vehicle velocity and the sideslip angle based on the sliding mode observer, the extended Kalman filter (EKF) including Dugoff's tire model, and the scaled Kalman filter with model error compensator (SKFMEC) methods, respectively. The estimation performance of the monitoring systems is verified in field tests, and the experimental results demonstrate the effectiveness of this approach to provide accurate monitoring of lateral vehicle motion.

The longitudinal and lateral vehicle control techniques have been widely used in several active driver assistance systems. The adaptive cruise control, lane keeping assistant control, vehicle platooning and stop-and-go control are typical examples of the most important applications. In this study, a novel path planning method is proposed considering the driving environment such as road shape, ego vehicle and surrounding vehicles’ movement. The relative distance and velocity between the ego vehicle and surrounding vehicles are identified with respect to the predicted lane shape in front of the ego vehicle. Based on the identified information, the road shape and surrounding vehicles are mapped into the intensity image and the desired vector for the ego vehicle’s movement is determined by the maximum intensity density tracing method. The desired vehicle path is followed by the acceleration/deceleration control and the steering assist control, respectively. In order to evaluate the performance of the proposed system, simulations are conducted and compared with ACC systems.

For high-speed/high-accuracy position control of a gantry-moving-type linear motor, we propose a nonlinear adaptive controller including a synchronization algorithm. Linear motors are easily affected by force ripple, friction, and parameter variations because there is no mechanical transmission to reduce the effects of model uncertainties and external disturbances. Synchronization error is also caused by skew motion, model uncertainties, and force disturbance on each axis. Nonlinear effects such as friction and ripple force are estimated and compensated for. The synchronization algorithm is used to reduce the synchronous error of the two side pillars. The performance of the controller is evaluated via computer simulations.

For advanced electron beam lithography systems, a simulation tool for a two-axis nano stage is developed in this paper. The stage is equipped with piezo-actuators and flexure guides. Even if piezo-actuators are believed to be feasible for realizing nano scale motions, it is difficult to predict their characteristics due to their nonlinearities such as hysteresis and creep. In this paper, the nonlinear properties are modeled considering the input conditions. In detail, the hysteresis is described as a first order differential equation with 24 sets of the hysteresis parameters and the creep is modeled as a time-dependent logarithmic function with two sets of creep parameters. The characteristics of the flexure guides are analyzed using the finite element method and are embodied into a multi-body-dynamics simulation tool. The dynamic behavior of the simulation tool is compared with the experimental data.

Linear motors are easily affected by load disturbances, force ripples, friction, and parameter variations because there are no mechanical transmissions that can reduce the effects of model uncertainties and external disturbance. In this study, a nonlinear adaptive controller to achieve high-speed/high-accuracy position control of a two-axis linear motor is designed. The operation of this controller is based on a cross-coupling algorithm. Nonlinear effects such as friction and force ripples are estimated and compensated for. An enhanced cross-coupling algorithm is proposed for effectively improving the biaxial contour accuracy while achieving closed-loop stability. The proposed controller is evaluated by performing computer simulations.

Vehicle detection is a crucial issue for driver assistance system as well as for autonomous vehicle guidance function and it has to be performed with high reliability to avoid any potential collision. The vision-based vehicle detection systems are regarded promising for this purpose because they require little infrastructure on a highway. However, the feasibility of these systems in passenger car requires accurate and robust sensing performance. In this paper, a vehicle detection system using stereo vision sensors is developed. This system utilizes feature extraction, epipoplar constraint and feature matching in order to robustly detect the initial corresponding pairs. The proposed system can detect a leading vehicle in front and can estimate its position parameters such as the distance and heading angle. After the initial detection, the system executes the tracking algorithm for the vehicles in the lane. The proposed vehicle detection system is implemented on a passenger car and its performances are verified experimentally.

As active chassis controllers are becoming increasingly complex and sophisticated, the performance of these controllers is very sensitive to a vehicle's inertial parameters such as the vehicle mass and mass centre height. Also, these parameters vary with the number of passengers and with different load situations. In this paper, in order to determine the vehicle mass during arbitrary vehicle manoeuvring, an online estimation system is proposed considering available driving conditions, which includes two estimation algorithms. One is designed on the basis of the vehicle's longitudinal motion using the recursive least-squares method with the disturbance observer technique. The other is designed on the basis of the vehicle's lateral motion by estimating the lateral velocity. Then, two estimation algorithms are combined to extract the vehicle mass information using the recursive least-squares method. The performance of the proposed estimation system is demonstrated through experimental tests.

A unified observer with stochastic and deterministic robustness is developed in this paper so that an observer is less sensitive to both stochastic and deterministic uncertainties. For stochastic robustness, the norm of the observer gain and the lower bound of the observer decay rate are shown to be design factors which can minimize the upper bound of the estimation error variance. For deterministic robustness, the L 2 norm-based condition number of the observer eigenvector matrix is utilized to address robust estimation performance against deterministic uncertainties. In order to justify the proposed method, a graphical approach is first introduced, and then a multi-objective optimization problem including linear matrix inequality constraints is formulated to provide the unified robustness.

The linear motors are easily affected by load disturbance, ripple force, friction, and parameter variations because there is no mechanical transmission to reduce the effects of external disturbance. For high-speed/high-accuracy position control of a linear-motor-driven X—Y table, a non-linear adaptive controller including a cross-coupling algorithm is proposed in this paper. The non-linear effects such as friction and ripple force are estimated and compensated. The cross-coupling algorithm is adopted to reduce the contour error of the two-axial system. The performance of the proposed controller is evaluated through the computer simulations and experiments.