International Journal of Vehicle Autonomous Systems (8 papers in press)
Real-time automatic obstacle detection and alert system for driver assistance on Indian roads
by Sachin Sharma
Abstract: Road crashes have been a major problem in India in recent times. The occurrences have increased considerably owing to the influx of four-wheelers and two-wheelers on Indian roads. The numbers of road traffic collisions have also increased owing to the absence of automatic highway safety and alert systems on major roads connecting cities and towns. The interior roads connecting villages and towns have been instrumental in multiple animal-vehicle collisions. Although the figure is not too large compared with other causes of road-related injuries, they are significant in number. The associated numbers of fatalities and injuries are substantial too. Though numerous efforts have been in progress to solve and reduce the number of collisions, lack of practical applications and resources, along with quality analytical data (for training and testing) related to animal-vehicle collision, has impeded any major breakthrough in the scenario. In our current work, we have proposed and designed a system based on histogram research including oriented gradients and boosted cascade classifiers for automatic cow detection. The Indian cow has been the commonest obstacle compared with other animals on Indian roads. The distance between a cow and the vehicle is calculated, prompting an alert signal to notify the driver for applying brakes or undertaking any similar action. The method is implemented in OpenCV software and tested on various video clips involving cow movements in various scenarios. The proposed system has achieved an overall accuracy of 80% in terms of cow detection. The proposed system is a low-cost, highly reliable system which can easily be implemented in automobiles for detection of cows or any other animal after proper training and testing on the highway.
Keywords: animal detection system; automotive electronics; cascade classifier; computer vision; histogram of oriented gradient; intelligent highway safety; OpenCV; road accidents.
Three-axle commercial vehicle with enhanced functionality and steering redundancy
by Dan Williams
Abstract: This paper presents a new means of steering redundancy for autonomous three-axle vehicles that also cost effectively increases vehicle functionality. Functional enhancements to increase maneuverability and decrease tyre wear are already appreciated in niche vehicle markets and are reviewed in this work. Steering the rear axle to provide redundancy in event of a primary steering axle failure has recently been suggested. This prior work is built upon to present a new three-axle vehicle configuration that improves maneuverability, increases payload capacity, and provides better redundant directional control while maintaining the tyre wear improvements existing in rear axle steer vehicles. Some of these same benefits could be achieved by steering the rear of a two-axle vehicle, but it is shown that the concept creates more value when applied to three-axle vehicles, thereby uniquely improving the value proposition for autonomous commercial vehicles.
Keywords: autonomous vehicles; commercial vehicles; vehicle dynamics; rear axle steer.
Stability control of a road vehicle considering model and parametric uncertainties
by Mazaher Mehdizadeh, Mahdi Soleymani, Amirhossein Abolmasoumi
Abstract: Parametric uncertainties due to unknown parameters of the model and unmodelled dynamics are two factors that degrade performance of vehicle stability control systems. In this paper, a modified sliding mode control system is proposed to enhance the handling performance of a passenger car via differential braking in the presence of road and model uncertainties. The proposed control system consists of two control loops and stabilises the vehicle via minimising the desired and achieved yaw rates. The inner loop, which uses a sliding mode controller, is responsible for calculating the required yaw moment. The outer control loop, which plays the role of supervisor controller, determines the braking pressure at each wheel. The controller is designed based on the reduced bicycle model and is implemented in a virtual prototype model with sufficient degrees of freedom. The simulation results establish the ability of the proposed control system in vehicle rapid stabilising under a severe manoeuvre. Moreover, the robustness of the controller in the presence of variations of vehicle mass and road friction coefficient was proved.
Keywords: stability control; uncertainties; sliding mode control; fuzzy control; differential braking.
Adaptive backstepping discrete-time control for a full-car active suspension
by Toshio Yoshimura
Abstract: This paper presents a simplified adaptive backstepping control (ABC) in the design of an active suspension for a full-car model with unknown external disturbances. It is assumed that the full-car models are described by an uncertain discrete-time state equation, and that the observation of the states is taken with measurement noises. The proposed ABC is designed in a simplified structure by removing the repeated heavy computation of nonlinear functions, and the design parameters are selected by using an appropriate Lyapunov function. The unmeasurable states and uncertainties for the uncertain discrete-time state equations are estimated by using the simplified weighted least squares estimator. The simulation experiment indicates that the proposed ABC is suitable for use in the design of the active suspension for the full-car model because it suppresses the movement of the car body and the suspension travel, and improves the ride comfort of passengers.
Keywords: adaptive backstepping control; virtual control; full-car model; discrete-time system; non-physical model; simplified weighted least squares estimator; active suspension; unknown road disturbance; Lyapunov function; ride comfort; simulation.
Identification of the driving style for the adaptation of assistance systems
by Görkem Büyükyildiz, Olivier Pion, Christoph Hildebrandt, Martin Sedlmayr, Roman Henze, Ferit Küçükay
Abstract: Exact and reliable knowledge of specific driver characteristics, such as driving style, open loop and closed loop behaviour, performance level and age, is important for the adaptation of assistance systems to the driver. The increased safety and comfort based on this knowledge can improve customer acceptance. This study presents a method of identifying specific driver characteristics, i.e. a personal fingerprint. This can be used to draw conclusions about the driving style, age and performance of the driver. To identify the fingerprint, the driver is classified based on the individual driver behaviour. For this purpose, objective parameters are defined from longitudinal and lateral control behaviour. In addition, steering and lane keeping behaviour are also analysed. The method of identifying the driving style, which is defined as the first component of the fingerprint, and its implementation into the vehicle using a driving style identifier are the main focuses of this paper. For this purpose, an algorithm to identify the driving style based on driving dynamic data is extended with environmental sensor information. To improve the driving style identifier, lane camera and radar data is taken into account in addition to vehicle signals, such as velocity, longitudinal and lateral acceleration. The lane keeping and following behaviours of the different types of driver are therefore also integrated into the identification process. Several examples of the process of determining the objective parameters from longitudinal and lateral control behaviour are illustrated.
Keywords: driver behaviour modelling; driver monitoring; adaptation of assistance systems; driver state detection.
Estimation of vehicle sideslip angle and individual tyre-road forces based on tyre friction circle concept
by Hui Lu, Qingwei Liu, Yue Shi, Fan Yu
Abstract: The real time information of vehicle sideslip angle and tyre-road forces of individual wheels can help advanced vehicle chassis control systems to enhance vehicle handling, stability and safety. But in practice, these state variables are difficult to measure directly for technical and economic reasons. In order to esti-mate these states, this paper proposes an observer based on the Extended Kalman Filter (EKF) by using a 7-DOF vehicle model. According to the Dugoffs tyre model, the lateral force can be expressed by a function of the longitudinal force with the knowledge of tyre work condition. Based on this concept, the reference vehicle model is modified to identify the lateral forces of each braked wheel without the online information of vertical load and tyre-road friction coefficient. The simulation results indicate that the longitudinal and lateral forces of each wheel can be well estimated under combined cornering and braking conditions.
Keywords: tyre force estimation; vehicle sideslip angle estimation; tyre friction circle concept; extended Kalman filter.
Modelling and optimisation of active front wheel steering system control for armoured vehicle for firing disturbance rejection
by Mazuan Mansor, Khisbullah Hudha, Zulkiffli Abd Kadir, Noor Hafizah Amer, Vimal Rau Aparow
Abstract: While firing on the move, the handling performance of an armoured vehicle is affected, thus causing it to lose its directional stability. This is owing to an impulse force generated at the centre of the gun turret, which can produce an unwanted yaw moment at the centre of gravity of the armoured vehicle. Hence, in order to reject the unwanted yaw moment, a new hybrid control strategy known as neural-PI controller is introduced by combining a neural network system and a conventional PI controller. This active system, furthermore, is intended mainly to reduce the yaw rate and to enhance the handling performance by providing the steering correction angle to the conventional steering system. Moreover, this paper develops 14 DOF of armoured vehicle and 2 DOF of Pitman arm steering system. Also included in this paper is the dynamic equation of the planetary gear system used as the active front wheel steering actuator implemented in the vehicle model, which is derived by using the Euler-Lagrange method. Other than that, determination of the most suitable activation function to be implemented in the neural-PI controller is carried out and optimised by using the genetic algorithm method. The performance of the controller is further evaluated by comparing the conventional PI controller with the neural-PI controller implemented with different activation functions.
Keywords: 14 DOF; active front wheel steering; firing disturbance; neural network; genetic algorithm; activation function.
Global path planning for autonomous vehicles in off-road environments via an A-star algorithm
by Qinghe Liu, Lijun Zhao, Zhibin Tan, Wen Chen
Abstract: In order to solve the problem of global path planning for autonomous vehicles in off-road environments, an improved A-star path-searching algorithm that considers the vehicle powertrain and the fuel economy performance is proposed in this paper. First, we discuss the digital elevation model (DEM) matrix adopted to describe the off-road earth surface generally. Then, we define three important concepts regarding path planners on the basis of the DEM matrix. Second, we design a novel comprehensive cost function for the A-star algorithm with shortest path distance and minimum fuel consumption. Finally, the algorithm is simulated on a DEM through several different missions. The simulation results show that the proposed algorithm is effective and robust in finding the global path in complex terrains.
Keywords: A-star algorithms; path planning; autonomous vehicles; digital elevation model.