Search

This paper presents a real-time optimal motion planner algorithm for road vehicles. The method is based on a cubic spline trajectory planner which is able to plan a set of vehicle motions driving from a given initial state to a required final state. Maximal dynamical feasibility and passenger comfort are ensured by minimizing the lateral acceleration and tracking errors as the vehicle moves along the trajectory. Tracking of the planned motion is realized during planning and execution as well by separate longitudinal and lateral controllers. efficient implementation and small number of optimization variables enables real-time usage. The trajectory planner is first tested in a quasi real-time simulation environment and then under real working conditions at the dynamic platform of proving ground ZalaZone with a completely drive-by-wire Smart Fortwo. Measurement results are presented and analyzed in detail, and possible future research directions are mentioned.

Environment perception and situation awareness are keystones for autonomous road vehicles. The problem of maneuver classification for road vehicles in the context of multi-model state estimation under model uncertainty is addressed in this paper. The conventional approach is to define different motion models that match the desired type of movements. In this work we used a single motion model as a starting point and applied constraints to construct such filters that are fine tuned for the predefined maneuvers. The estimation is carried out in the interacting multiple model framework, where the elemental filters are constrained Kalman filters. To capture the characteristics of the considered maneuvers linear equality and non-equality state constraints were used. The performance of the proposed method is demonstrated in a simulation environment participating an observer and a maneuvering vehicle.

The paper presents a small-scale electric vehicle framework for vehicle control education and research. The main goal of the project is to serve as a good experimental platform for the students on any level of vehicle mechatronics education. It offers wide range of possibilities for embedded system, control design and machine learning applications. The proposed system is a redesigned version of our former experimental vehicle framework. The project was scheduled for two main goals. One was to design and implement a Robot Operating System based vehicle with Ackermann steering. The other was to develop a platform, which can be highly integrated into vehicle control education. In the paper the system architecture, the sensors and the control units are detailed, furthermore the educational benefits and some use cases are presented.

The paper presents a microscopic highway simulation model, built as an environment for the development of different machine learning based autonomous vehicle controllers. The environment is based on the popular OpenAI Gym framework, hence it can be easily integrated into multiple projects. The traffic flow is operated by classic microscopic models, while the agent’s vehicle uses a rigid kinematic single-track model, with either continuous or discrete action spaces. The environment also provides a simple high-level sensor model, where the state of the agent and its surroundings are part of the observation. To aid the learning process, multiple reward functions are also provided.

The paper presents a reinforcement learning based solution for the control design problem of a gearbox actuator. The system is operated by an electro-pneumatic, three-state, floating piston cylinder. Besides the primary goals of positioning the piston, the nonlinear system’s quality objectives are to minimize switching time and overshoot. The control strategy based on the measurable parameters of the system is realized by a dense feedforward neural network. With the utilization of the policy based reinforcement learning architecture, the learning agent develops the optimal strategy for fast and smooth switching, under different and changing conditions.

The Bernoulli filter (BF) in the interacting multiple model (IMM) framework is proposed for detecting and tracking a maneuvering target. The BF is implemented as a particle filter and embedded in the IMM structure. The communication between the IMM and the BF is achieved through a Gaussian layer. Particles are drawn from the mixture densities at each recursion and the model likelihoods are computed from the filter innovations. Simulations show that the proposed filter outperforms the single model variant and it can effectively choose the correct motion model and estimate the state of the tracked object.

Cooperative multi-model state estimation is considered in road traffic scenario. Our goal is to detect and track a maneuvering object based on vehicular radar measurements. To handle the motion model switching of the navigating object we used the interacting Multiple Model (IMM) estimator with the Bernoulli filter (BF). We modified a recently derived IMM-BF, realized in a Gaussian mixture approximation to fulfil the requirements of a multi-sensor object tracking estimator with moving sensors. Our proposed method is implemented as a Sequential Monte Carlo estimator that allows nonlinear motion and sensor models and is capable of functioning effectively in a cooperative manner.

The aim of this paper is to present a local trajectory planning method based on nonlinear optimization that is able to generate a dynamically feasible, comfortable and customizable trajectory for highly automated road vehicles. The presented algorithm is able to consider the nonholonomic dynamics of wheeled vehicles and ensures the dynamical feasibility of the planned trajectory by the model-based prediction of the vehicle’s motion. The behavior of the vehicle is simulated with closed loop trajectory tracking control which allows to generate not only the trajectory of the vehicle but also the reference signal inputs for the controllers. The direct planning of the reference signals enables the vehicle to run exactly on the generated trajectory and eliminates the delays related to the inertia of the system.

Multi-sensor object detection and tracking on a highway scene with radar measurements is presented. The estimation algorithm is the random finite set based Bernoulli filter, working in the Bayesian framework. The recursion for calculating the Bayes estimation is implemented as a particle filter. A method is presented for calculating the likelihoods, suitable for particle filtering performed with moving sensors, assuming additive Gaussian measurement noise. In our approach, for calculating the posterior estimate of the object state, the measurement likelihoods are computed in the state space, instead of the measurement space, by mapping each measurement to the global coordinate system. The map consists of a nonlinear and an affine part. While the affine transformation trivially preserves the Gaussian nature, the nonlinear is well-proven to be approximated as affine too. This approach allows the particles to be drawn directly from the state space, hence the evaluation of the measurement model is not needed, which saves computational power.

The paper presents the development of an experimental vehicle framework for testing autonomous vehicle functions. The development of system serves primarily the education and the scientific work of the students. It is a good experience for the students on any level of mechatronic education because of its diverse possibilities for mechanical engineering, embedded systems, or control design. Overall system architecture, sensors, actuators and control solutions are outlined in this paper along with the development of a low-cost 3-D LIDAR sensor and some Simultaneous Localization and Mapping (SLAM) experiments.

For future’s highly automated road vehicles, dynamically feasible, comfortable, and customizable trajectories must be planned in order to ensure the maximum level of road safety and passenger satisfaction. To fulfil these requirements, a constrained nonlinear optimization based trajectory planning method was developed, which is generating the trajectories by the model-based prediction of the vehicle’s motion. The evaluation of vehicle dynamics in closed loop with the trajectory tracking control can also allow the direct planning of controller reference signals, which is unique in literature. The aim of this paper is to investigate the sensitivity and robustness of the planner method, based on simulation results, taking special account of the impact of deviations from normal operating circumstances such as rapid changes in road surface. These examinations are almost always missing from current works, although they are indispensable for safe and feasible motion planning under wide variety of road and environmental conditions. Firstly, the developed trajectory planning algorithm is described. Then the performance of the method is examined in terms of changes in the parameter values of the applied model of vehicle dynamics, tire-road contact, and the optimization problem setup. Finally, the performance, sensitivity, and robustness of the presented trajectory planning algorithm is evaluated based on the simulation results, and the conclusions are summarized.

The reduction of energy consumption has gained significance in the past years as energy prices have been continuously increasing. Advances in railway telematics and the large amount of data obtained from train services enable the development of methods that are capable of further improving energy efficiency through the evaluation, control or prediction of energy utilization of the railways. The paper presents the results of a statistical analysis in which the rate of the possible energy savings is calculated using on-line telemonitoring system data. During the analysis one month real train data was processed from the MÁV Electronic Logbook System, narrowing the data to the electric locomotives.

The paper presents the performance evaluation of a multi-sensor object detection algorithm applied in traffic situation. The chosen data fusion and estimation procedure is the Bernoulli particle filter, which is ideal for cooperative object detection, as it can handle the varying number of sensor measurement and object appearance-disappearance too. The simulated detectors are automotive radar sensors, capable of measuring object speed, distance and bearing. The filter performance is examined along four aspects: number of particles used, simulation timestep, sensor noise and object motion model noise. The applied noise models represent the quality of the equipped sensors and our lack of knowledge on object motion, therefore the robustness of the filter can be examined by varying these parameters. The amount of particles and the used timestep effect the demanded computational power and also the estimation error, hence the scaling of the algorithm can be measured. In order to obtain a statistically meaningful result, a batch of 50 runs for every parameter set were used.

Controllers for autonomous vehicle strings usually consist of local feedback controllers responsible for performing acceleration demands and a common control law aimed at the string stability of the entire platoon. Based on analysis of robust peak-to-peak performance and experimental verification, it is shown in the paper that satisfactory spacing performance can be achieved without accurate knowledge on the engine/brake/gear systems, only by using a common control law that relies on inter-vehicle communication and the commercial on-board services of today's heavy vehicles.