This article proposes the optimal discrete-time integral sliding-mode control for unmatched external signal compensation in a lane-keeping system on curved roads. Since curved roads appear as an external signal in a lateral vehicle motion model, improving the tracking performance with external signal compensation is necessary. To tackle the problem, we solve the optimization problem of obtaining the optimal sliding surface that minimizes the$\mathcal {H}_{2}$norm of the transfer matrix from the external signal to the system state of interest. As a result, robust performance in the sense of$\mathcal {H}_{2}$is guaranteed in the proposed framework. We provide a theorem to prove the$\mathcal {H}_{2}$performance and present the design process. The proposed method is applied to lateral vehicle control to validate the effectiveness via numerical simulation and real-world experiments. A comparative study confirmed that the tracking performance of the proposed algorithm outperforms those of other methods on curved roads.

The potential functions (PFs) have generally shown good performances in real-time path planning with computation efficiency conforming to the requirements of lower control systems in autonomous driving. However, several inherent limitations exist in using the PFs, including a local minimum in specific scenarios and no passage between closely spaced obstacles. Recent studies have focused on conventional scenarios where PFs are assumed to work normally, without malfunctioning, occurring during perilous situations. Therefore, we propose a specific safety tunnel (ST)-based model predictive controller (MPC) combined with PFs (PF-STMPC) to handle path-planning in extreme-emergency traffic scenarios (e.g., emergency braking and lane-changing obstacles). To further guarantee driving safety, we improve PFs with the responsibility-sensitive safety (RSS) model that accurately calculates the minimum safe longitudinal and lateral distances. Furthermore, a sigmoid-based ST is designed for emergency navigation if the PFs fail to plan a safe path due to the aforementioned inherent limitations, enabling the controller with planning functionality if necessary. The ST is embedded in the MPC-based tracking controller as a safe constraint sensitive to surrounding environments (e.g., road structure and obstacles). The proposed PF-STMPC was co-simulated using MATLAB/Simulink and CarSim Simulator under the constant speed condition. Compared with the state-of-the-art method, the proposed method demonstrated better performance in finding a safe path and eliminating severe yawing of the ego-vehicle (82.8% less in sideslip yawing amplitude and 57.7% shorter in the oscillation period of yaw angle) when facing traffic emergencies.

In this article, we present an innovative approach, i.e., horizonwise model-predictive control (H-MPC), to solve the model-predictive control (MPC) problem of a linear time-varying (LTV) system. In H-MPC, we regard the time-varying parameters as time invariant within the prediction horizon. To solve the MPC problem of the time-varying system, the decision variable is decomposed into two terms: one for linear time-invariant optimization and the other for compensating LTV uncertainties with an introduction to a uniform compensation condition. The proposed H-MPC solves the time-varying problem by removing the uncertainty due to the future parameter variations within the horizon and by updating the time-invariant MPC at each sampling time. To validate the usefulness of the proposed H-MPC, it is applied to lane tracking control for an autonomous driving vehicle. From a comparative study of the H-MPC and conventional MPCs in lane tracking control, it is confirmed that the proposed H-MPC has a competitive performance compared to LTV-MPC despite its much simpler structure.

This paper proposes an object collision point estimation scheme by developing a new data fusion method in a multi-radar network environment. In order to reduce radar’s estimation error due to measurement uncertainty, we first design radar accuracy models determined by the position of each object. Then, an interacting multiple model (IMM) filter based on occupancy zones is designed for accurate object estimation. For a multi-radar network’s object estimation, we also design a radar data fusion method using the estimated object information through the IMM instead of the object estimation information given by the radars. A collision point identification problem, where multiple sensors calculate the different vehicle surface points of the same object, is solved by developing the data fusion method to estimate the object surface’s collision point closest to the ego vehicle center. The utility of the proposed scheme was validated through a scenario-based object estimation experiment. We confirmed that the proposed data fusion method produced substantially improved error distributions over conventional methods.

In this paper, a robust$\mathcal {H}_{2}$controller based on Linear Parameter Varying (LPV) model with scheduling variable reduction is applied for a lane-keeping system. On a curved road section, varying longitudinal speed and roll motion lead to multiple parameter variations in the lateral vehicle dynamics. To trade-off between the complexity of the multiple scheduling variables and the accuracy of the LPV model, an order reducing method by Autoencoder (AE) is performed to obtain a reduced model and diminishes the conservatism of LPV-based controller design. Further, to ensure the continuous convex set membership of the reduced scheduling variables in online test scenarios, the Lipschitz constant of the offline trained neural network (NN) is tightly estimated by solving a Semidefinite Program (SDP). The convex set of local Linear Time-Invariant (LTI) controllers is designed according to the estimated Lipschitz bound of the trained NN. The LPV-based robust$\mathcal {H}_{2}$feedback controller is then designed by solving a set of Linear Matrix Inequalities (LMIs). Numerical simulations with full vehicle dynamics from CarSim are given to demonstrate the performance of the proposed system on various test roads. The continuous finding of the reduced scheduling variable membership is guaranteed in the online tests, and the effectiveness of the proposed method is confirmed with the mean value of lateral offset error reduced by about 50% compared with LTI-based$\mathcal {H}_{2}$controllers.

This paper proposes a collision detection system to detect whether ego and target vehicles collide when both vehicles change from their lanes to the same lane. Although it is essential to predict this kind of collision for the active safety system, there is little literature on the case study. This paper presents the collision detection method using a Convolution Neural Network (CNN) consisting of four classes to predict collision risk on multi-lanes road conditions. The CNN is formed on stacked Occupancy Grid Maps (OGMs) based on point cloud data of the LiDAR and Radar sensors with in-vehicle sensor data for spatio-temporal information between vehicles. Further, we apply the open set recognition concept to the network to consider a conservative collision detection. The experimental results show the feasibility of the proposed collision detection system and the conservative decision about the confusing situation.

In a dynamic lateral motion model with a look-ahead distance, curved and winding roads appear as an unmatched disturbance to a lane-tracking control system. In this article, an innovative vehicular robust lane-tracking control scheme is proposed to compensate for the problematic winding road disturbance (WRD) that always appears unless the road has zero curvature. We propose a design for a WRD compensator (WRDC) to substantially improve the lane-tracking control performance on a winding road in the presence of the WRD. The WRDC redesigns the state-space reference, rather than calculating the additional control input. We show that the proposed WRDC ensures that the projected tracking error decreases to zero. To validate the usefulness of the robust lane-tracking control, the WRDC is compared to other conventional methods. We observe that the proposed WRDC is not only robust against road curvature variation, but also outperforms the lane-tracking control performance of other popular methods.

In this article, a novel passivity-based control strategy is proposed for the exponentially stable tracking controller design of static synchronous compensator (STATCOM) system, which is a single input and single output. The STATCOM is not an input-affine system but a special port-controlled Hamiltonian system form. Hence, it is regularized by using a dynamic extension algorithm so that the proposed tracking control strategy is designed in an input-output linearization framework with a bounded solution to the driven zero dynamics equation. The proposed control strategy is proposed with consideration of the performance and stability of the input-output linearized dynamics. Simulation results show that the proposed control strategy improves the transient performance of the system compared to the previous results even in the lightly damped operating range.

In this paper, we present an innovative approach to steering control, based on torque overlay, for the smooth and efficient transfer of the steering control from an autonomous driving system to the driver, and to control the pinion angle of the rack-mounted electric power steering for lateral control of an autonomous vehicle. The novelty of our approach lies in the formulation of hybrid impedance control, which employs steering control for both lateral and impedance controls. The proposed method, nonlinear hybrid impedance controller, consists of desired impedance pinion angle generator and a super-twisting sliding mode controller for pinion angle tracking. The desired impedance transfer function of the driver's torque to the steering wheel angle reflects the driver's torque to the steering control. The desired impedance pinion angle is generated using the desired impedance transfer function and the desired pinion angle derived by the lateral controller. The sliding mode control method is designed, based on a super-twisting algorithm for the pinion angle, to track the desired impedance pinion angle. Consequently, without the driver's torque, the proposed method is activated as a pinion angle tracking controller for lateral control of an autonomous vehicle. Moreover, with the driver's torque, the proposed method is activated as an impedance controller.

In this paper, we introduce a new deep learning architecture for camera and Lidar sensor fusion. The proposed scheme performs 2D object detection using the RGB camera image and the depth, height, and intensity images generated by projecting the 3D Lidar point cloud into camera image plane. The proposed object detector consists of two convolutional neural networks (CNNs) that process the RGB and Lidar images separately as well as the fusion network that combines the feature maps produced at the intermediate layers of the CNNs. We aim to develop a robust object detector that maintains good object detection accuracy even when the quality of the sensor signals is degraded for object detection. Towards this end, we devise the gated fusion unit (GFU) that adjusts the contribution of the feature maps generated by two CNN structures via gating mechanism. Using the GFU, the proposed object detector can fuse the high level feature maps drawn from two modalities with appropriate weights to achieve robust performance. Experiments conducted on the challenging KITTI benchmark show that the proposed camera and Lidar fusion network outperforms the conventional sensor fusion methods even when either of the camera and Lidar sensor signals is corrupted by missing data, occlusion, noise, and illumination change.

In this paper, we propose a deep learning based vehicle trajectory prediction technique which can generate the future trajectory sequence of surrounding vehicles in real time. We employ the encoder-decoder architecture which analyzes the pattern underlying in the past trajectory using the long short-term memory (LSTM) based encoder and generates the future trajectory sequence using the LSTM based decoder. This structure produces the K most likely trajectory candidates over occupancy grid map by employing the beam search technique which keeps the K locally best candidates from the decoder output. The experiments conducted on highway traffic scenarios show that the prediction accuracy of the proposed method is significantly higher than the conventional trajectory prediction techniques.

In this paper, we propose a vehicle path prediction employing yaw acceleration for adaptive cruise control (ACC). First, a path prediction method employing yaw acceleration is proposed to improve the path prediction performance of ego vehicles. In the proposed method, the vehicle path is predicted by using a clothoidal cubic polynomial curve model, and for this purpose, the curvature rate, yaw rate, and longitudinal velocity are required. The curvature rate can be mathematically obtained by differentiating the yaw rate without a camera sensor. To obtain the yaw acceleration from the noisy measured yaw rate, we derive the state-space model from the steer wheel angle for the yaw rate. Then, the KF is designed by using the state-space model to estimate the yaw acceleration. Second, a multirate longitudinal control method is proposed to improve the longitudinal control performance. The multirate KF employs the constant acceleration model in order to estimate the relative distance and velocity of the target vehicle at a faster sampling rate. Then, the desired acceleration is achieved to maintain a safe headway distance or velocity by means of the longitudinal controller. Consequently, the whole ACC system operates with a faster sampling rate so that multirate control scheme reduces ripples in both the relative longitudinal distance and the desired acceleration. The performance of the proposed method was evaluated via simulations and experiments.

In this paper, we propose a flatness based controller which tracks the desired angle of an electric power steering system with augmented observer and a state feedback controller. Unlike the torque-overlayed backstepping method in a lane-keeping system, using the known nominal model parameters results in better tracking performance then the previous method. This method can be implemented as a torque-overlay control as the backstepping controller. From numerical simulations, we observed that the flatness-based steering wheel controller outperforms the backstepping controller.

It is known that the kinematic model based motion control is robust against unknown vehicle parameters variation. Recently we reported that lane keeping system (LKS) with look-ahead distance using the kinematic vehicle lateral motion model is feasible and its performance is compatible with a dynamic vehicle lateral motion model using look-ahead distance under a limited condition in highway driving. In this paper, we developed a kinematic vehicle motion model with road friction coefficient estimation. The adaptive kinematic vehicle motion based lateral controller requires no complex tuning process. The proposed method provides improved LKS performance over the previous method. Control performance of model was validated via computational simulation results with CarSim and MATLAB/Simulink.

This paper presents a new direct active and reactive power control (DPC) controller scheme for a three-phase grid connected voltage source inverter (VSI) based on passivity viewpoint. The proposed method is designed in the framework of port-controlled Hamiltonian system. The proposed method is made up of two parts, one thing is a feedfoward part and the other is a feedback part. The feedforward part is calculated from the model dynamics of VSI. The feedback part is designed to compensate for system uncertainty. The proposed method is verified with simulation and experiment. The simulation uses MATLAB/Simulink and PLECS, and the experiment uses hardware in the loop (HIL) with TI TMS320F28335 DSP. The simulation and experiment results compared with those using direct power control-sliding mode control (DPC-SMC) for the grid connected VSI. The proposed method gives less total harmonic distortions, faster transient response.

This paper presents gain scheduling H2controller for speed control of permanent magnet synchronous motors (PMSMs) in the sense of linear parameter varying (LPV) synthesis. First, the nonlinear dynamics of a PMSM is transformed to a LPV system which has state-dependent varying parameters. The proposed method is composed of torque modulation, commutation scheme and H2controller based on parameter-dependent system. The stability of closed-loop system is proven by Lyapunov theorem. The proposed method has not only fast velocity tracking performance but also reduced energy consumption. Finally, the effectiveness of the proposed method was validated with experimental results.

In this paper, we proposed nonlinear adaptive speed controller for permanent magnet synchronous motors in the presence of unbalanced phase winding resistances. An adaptive control is designed to compensate for the resistance unbalance in a-b-c frame dynamics. Estimated neutral voltage which is non-zero due to the unbalance is used for the adaptive control. It is proven using singular perturbation theory that the speed tracking error is globally asymptotically stable. Experiment with hardware-in-the-loop simulation (HILS) was performed for validating performance of the proposed method. We obtained that speed ripple was reduced with the proposed method.

In this paper, we are interested in the situation of the ego vehicle tracking the previous vehicle's GPS waypoint on highway. Even if waypoints are irregular, in order to improve the tracking performance and steering performance of the GPS waypoint tracking, curve fitting and model predictive control have been applied. The improvement of the performance of the waypoint tracking was validated via computational experimental results.

In this paper, we propose kinematic vehicle lateral motion model based lane keeping system considering look-ahead distance. The state-space model based on the kinematic vehicle lateral motion model is derived and we design the lane keeping system(LKS) based on the kinematic model. The kinematic model based LKS is robust against unknown vehicle parameters variation. Furthermore, to consider look-ahead distance in the kinematic vehicle lateral motion model, we designed output measurement matrix using clothoidal constraints. Hence, we can control the vehicle at look-ahead distance like human driver. Control performance of each model was validated via computational simulation results with CarSim and MAT-LAB/Simulink, and experimental results with electric power steering system equipped with an Autobox from dSPACE.

In this paper, we propose a new vehicle trajectory prediction algorithm for adaptive cruise control (ACC). When vehicle trajectory prediction is not precise enough, it is possible for a neighboring vehicle to be detected as a target. Thus, we propose a new method using both yaw rate and curvature rate to precisely predict vehicle trajectory and to resolve an undesirable case in ACC system. The proposed method was validated via CarSim and MATLAB/Simulink. Also, we validated the proposed method via experimental results with a test vehicle on highway system for the practicality.

In this paper, we propose a new lane estimation method to resolve poor detection performance of a camera system using the relative movement between a vehicle and lanes. In implementation of lane keep system (LKS) or lane change control (LXC), it is necessary to build a robust sensing system in obtaining the road information. When reliable road information is not available, we need virtual lane information to control the steering system. Thus we propose a predictive virtual lane (PVL) method to improve lane detection performance using the relative movement between a vehicle and lanes. The proposed PVL enhances the control performance of LKS/LXC for a while even for the detection failures of one side and/or both side lane marks. The performance of the proposed method was evaluated via simulations implemented with CarSim and Matlab/Simulink. Its performance was also validated with a test vehicle on highway system.

A multirate steering control scheme is developed for autonomous vehicle lateral maneuvering. The proposed scheme consists of a multirate extended Kalman filter and a state feedback control. The multirate extended Kalman filter is to estimate the vehicle states at a fast rate of the car ECU using multirate sensing - a slow vision-based lane detection by a camera and fast motion detection by inertia sensors. A method to design the multirate decentralized extended Kalman filter is presented. Through application results, the proposed multirate steering control scheme is shown to exhibit significantly improved control performance.