Abdelkader El Kamel

A safety spacing policy and its impact on highway traffic flow

This paper proposes a Safety Spacing Policy (SSP) that can ensure safe operation while at the same time improving traffic flow for Adaptive Cruise Control (ACC) system. The proposed SSP, a nonlinear function of vehicle velocity, uses both the information of vehicles state and braking capacity to adjust the position and velocity of the controlled vehicle. String stability, traffic flow stability and traffic capacity of the proposed policy are analyzed. The SSP can ensure vehicle string stability and it can also yield stable traffic flow and higher traffic capacity that are superior to the traditional Constant Time Gap (CTG) system. Especially in the high-density traffic conditions, the proposed new policy can provide higher traffic capacity to relief traffic congestion while the CTG policy fails to do so. In addition, traffic simulations show that SSP can ensure vehicle safety in hard brake and acceleration scenarios. The proposed new spacing policy is a promising alternative to the traditional CTG policy.

Coordinated throttle and brake fuzzy controller design for vehicle following

The coordinate throttle and brake control is one of the main challenges of vehicle longitudinal control. The main difficulties are due to the complexity and uncertainty of the vehicle dynamics, and the necessity of the switching between throttles and brakes. A hierarchical control system is proposed, in which an upper level controller determines the desired speed/acceleration for the controlled vehicle, while a throttle/brake fuzzy control system is proposed as the lower level controller. To coordinate the two speed actuators, a logic switch is designed. The simulation results show that the automated vehicles equipped with the proposed longitudinal control system have a good speed/spacing tracking capacity, and they are also adaptive to the disturbances of velocity variations and road grades. In addition, the operations of the throttle and the brake are smooth, which means the driving comfort is ensured.

Modeling intelligent vehicle agent in virtual reality traffic simulation system

Microscopic behaviors in traffic are complex and the traffic information is heterogeneous. However, current research on virtual reality traffic simulation pays little attention on modeling microscopic behavior with the consideration of the human factor. An intelligent vehicle agent model is proposed to better understandings of human factor in virtual reality traffic simulation system. The intelligent vehicle agent is a comprehensive model of 3D virtual vehicle and combines perception, motion, internal properties and cognition modules. Then, car following behavior model is introduced in detail, which is an application case to explain the function of intelligent vehicle agent. Car following behavior of virtual vehicle is realized by the coordination mechanism of agent-based multi-controller, which is determined by predetermined perceptual threshold levels. The preliminary results show the effectiveness of our method to establish the intelligent vehicle agent model and to simulate car following behaviors in virtual reality traffic simulation system.

V2X-Based Decentralized Cooperative Adaptive Cruise Control in the Vicinity of Intersections

Cooperative driving with V2X communication is widely researched due to its considerable potential to improve the safety and efficiency of road transportation systems. In this paper, a decentralized cooperative adaptive cruise control algorithm using V2X for vehicles in the vicinity of intersections (CACC-VI) is proposed. This algorithm is designed to improve the throughput of intersection by reorganizing the vehicle platoons around it, in consideration of safety, fuel consumption, speed limit, heterogeneous features of vehicles, and passenger comfort. Within a platoon, vehicles try to find the optimal control input by a distributed particle swarm optimization (PSO) algorithm, in order to reduce tracking errors, while respecting different constraints. A concept of opportunity space is proposed to facilitate platoon reorganization, in which a subplatoon or an individual vehicle can choose to accelerate to join in the preceding platoon or to decelerate to depart from the current one. The main idea is to make full use of the road capacity and to distribute it to most vehicles that are capable to find an accelerating trajectory to get through the intersection within a limited period. The originality of our algorithm is the introduction of a novel application of V2X communication to make the traffic more intelligent, in terms of safety, time saving, and environment friendly.

Multi-agent based modeling and simulation of microscopic traffic in virtual reality system

Abstract Traffic simulation in virtual reality system plays an important part in the research of microscopic traffic behavior, but developing the traffic simulation is a difficult work because of its inherent complexity. This article focuses on the modeling and simulation of microscopic traffic behavior in virtual reality system using multi-agent technology, a hierarchical modular modeling methodology and distributed simulation. Besides, the dynamic features of the real world have been considered in the simulation system in order to improve the microscopic traffic analysis. First, the multi-agent based system framework is designed and analyzed. Then, the environment agent and the intelligent vehicle agent are presented for the simulation of interaction between vehicles and environment, especially the road geometry and wind affect on the vehicle. Finally, the application results are presented to show the feasibility of the proposed method.

Lateral Control of Autonomous Vehicles Using Multi-Model and Fuzzy Approaches

Abstract This paper discusses the fundamental issues in vehicle lateral motion control within Automated Highway Systems. Lane keeping control and lane changing control are concerned. One major challenge for lateral controller design is the smoothness and robustness under different vehicle conditions. A multi-model fuzzy controller, which includes four local controllers, is then proposed for both the operations of lane keeping and lane changing. These four local controllers are set up for four vehicle speed regions respectively. A fusion block is then introduced to ensure smooth and accurate transition between the different local controllers. The proposed controller inherits the advantages of both the multi-model control and fuzzy control. Simulations show that it could get good performances in the whole range of operation speed, and could also repel the system uncertainties such as changes in vehicle load, movement inertia and wheel cornering stiffness. Furthermore, the calculation procedure of the proposed controller is not complex, and is rather rapid. It appears a promising control algorithm for realtime embedded applications.

Neural inverse reinforcement learning in autonomous navigation

Abstract Designing intelligent and robust autonomous navigation systems remains a great challenge in mobile robotics. Inverse reinforcement learning (IRL) offers an efficient learning technique from expert demonstrations to teach robots how to perform specific tasks without manually specifying the reward function. Most of existing IRL algorithms assume the expert policy to be optimal and deterministic, and are applied to experiments with relatively small-size state spaces. However, in autonomous navigation tasks, the state spaces are frequently large and demonstrations can hardly visit all the states. Meanwhile the expert policy may be non-optimal and stochastic. In this paper, we focus on IRL with large-scale and high-dimensional state spaces by introducing the neural network to generalize the expert’s behaviors to unvisited regions of the state space and an explicit policy representation is easily expressed by neural network, even for the stochastic expert policy. An efficient and convenient algorithm, Neural Inverse Reinforcement Learning (NIRL), is proposed. Experimental results on simulated autonomous navigation tasks show that a mobile robot using our approach can successfully navigate to the target position without colliding with unpredicted obstacles, largely reduce the learning time, and has a good generalization performance on undemonstrated states. Hence prove the robot intelligence of autonomous navigation transplanted from limited demonstrations to completely unknown tasks.

Modeling overtaking behavior in virtual reality traffic simulation system

Overtaking behavior in traffic is complex and the traffic information is heterogeneous, current researches on virtual reality traffic simulation have paid little attention on modeling overtaking behavior, considering the different traffic situation. In this paper, an intelligent vehicle model is proposed to better understandings of traffic situation and to assist overtaking behaviors analysis in traffic simulation. Then overtaking behavior model based on the intelligent vehicle is introduced in detail, the lane changing behavior model is analyzed because it is the base of overtaking behavior. Besides, overtaking behavior is realized by the coordination mechanism of agent-based multi-controller, which incorporates different traffic situation to explore overtaking behavioral mechanism in traffic. The preliminary results show the effectiveness of our method to simulate overtaking behavior in virtual reality traffic simulation system.

Real time indoor intelligent navigation system inside hypermarkets

Abstract This paper introduce a real time indoor navigation system dedicated to assist the handicapped persons inside hypermarkets. This system is designed to guide them during their shopping to show them, in real time, the way to find their items. The particularity of this system is that it represents a GPS indoor system which determines the users location in the hypermarket and computes the shortest path to pick up all items. In this paper, we will focus on the itinerary optimization problem. For this, we will introduce a method to compute the distance between any two items in the hypermarket and find the shortest way.

Integrated Longitudinal and Lateral Control System Design for Autonomous Vehicles

Abstract Abstract This paper presents the design of an integrated longitudinal and lateral controller for autonomous vehicles. Firstly, the problem of longitudinal controller design is to concentrate on the design of spacing policy and its associated control law. A Safety Spacing Policy (SSP) is proposed, and a sliding mode controller is designed. String stability as well as traffic flow stability issues of the proposed SSP are studied. Its shown that the SSP can provide stable string operation and stable traffic flow under certain conditions. Secondly, a multi-model fuzzy controller is proposed to cope with the vehicle lateral control. This multi-model fuzzy controller contains four local fuzzy controllers which correspond to four velocity intervals. In order to avoid the undesired peaks which may appear during the commutation of the different local controllers, a fusion block based on a fuzzy type strategy is designed. Finally, the integration of the longitudinal and lateral controllers is presented, and the simulation results show that the proposed controller has good performances under different scenarios.