This report will present the results of the initial study to develop an accurate working model of the 40 foot New Flyer Bus powered by the Cummins C8.3+ 280G CNG engine (Figure 1.1). The main focus of the study is the modeling of the vehicle dynamics using step input acceleration data. Other parts of the research include gear shifting and torque production from the CNG engine, all of which are required to produce a completely simulationcapable bus model.

The aim of this report is to study issues concerning fault detection and identification in vehicle follower control. The report discusses various potential fault modes among sensors, and actuators used in the vehicle following experiments in cars. It deals with issues of fault detection of sensors and actuators for vehicle following applications.

One of the main goals of an Automated Highway System environment is to increase the throughput of vehicles traveling on the highway. By moving vehicles in a platoon (a group of tightly spaced vehicles), the traffic flow capacity can be greatly increased. The control law developed for vehicles to safely travel in a platoon is dependent on the lead and preceding vehicle's velocity and acceleration profiles. This information guarantees string stability (i.e. spacing errors between vehicles do not increase down the chain of vehicles). These profiles are transmitted to the vehicle via wireless communication. Unfortunately, a perfect wireless communication does not exist. In this paper, the effects of various communication delays on string stability will be analyzed. The concept of platooning in an Automated Highway System (AHS) allows a group of vehicles to share information across a wireless local area network (LAN). This sharing of information allows vehicles belonging to the same platoon to maintain a smaller inter-vehicular spacing that would otherwise be possible. Of course, once these platoon/LANs exist on the AHS, a method must exist to add vehicles to a platoon and also to remove vehicles from a platoon. This report also develops handshaking protocols that allow the LANs associated with each platoon to reconfigure themselves in response to any physical changes to the composition of the platoon. Since the LANs operate over a wireless communication system, these protocols are designed to be robust towards packet losses, as well as satisfying certain safety and liveness conditions.

The number and quality of sensors available for both on-board vehicle and infrastructure-based sensing is increasing while the cost of these sensors is rapidly decreasing. On-board vehicle sensors can be utilized for both individual vehicle safety as well as automated vehicle control. It is becoming imperative that "fusion" techniques be developed, i.e., methods to combine the wide variety of sensors available so that reliable and accurate information can be obtained even though individual sensors may fail or become extremely noisy. The process of sensor fusion combines multiple sensor measurements in order to provide an accurate and coherent view of the vehicle and its environment.

A new architecture for a mobile ad-hoc communication network is developed, and simulation results show that the proposed MAC protocol can enhance transmission success probability from 0 to 97%. Two novel applications of vehicle-to-vehicle networks are then developed and simulated. The first application is a \Cooperative Adaptive Cruise Con- trol" which uses communicated information to improve on ordinary cruise control systems. Communicating a \virtual brake light" between vehicles is shown to decrease the accelerations required to maintain safe following distances in cut-in and hard braking situations. The second application is a \Cooperative Estimation" algorithm. The idea behind this algorithm is that each vehicle on the road is potentially a driving condition \sensor." By combining data communicated from many vehicles on the roadway, the cooperative estimation algorithm is able to produce estimates of driving time and road condition that are significantly better than those that any one vehicle could produce on its own. While developing the cooperative road condition estimation algorithm, it is found that the problem's crux is at the vehicle level, so the friction estimation problem is then investigated at a vehicle/tire level through exten- sive experimentation. An exciting new type of \slip-based" road condition estimator is developed and experimentally shown to be able to distinguish between wet and dry roads using no dedicated road condition sensors sen- sors. In-depth literature review complements our own results and shows that \slip-based" estimators of the type developed may be very useful to help solve the AHS road condition estimation problem.

In order to provide reliable control algorithms on PATH automated highway system, knowledge of the transient aerodynamic forces is important from the point of the controlla- bility of the platoon and individual vehicles. In a passing maneuver, the platoon members may experience severe aerodynamic forces and yaw moment when another vehicle overtakes. Similar phenomenon can be seen in our daily life as one car passes the other but much more complicated flow field is expected when the interaction involves in a multi-vehicle platoon. It is the purpose of this study to quantify the unsteady aerodynamic forces and moment so as to assure the safety of the system.

This report deals with an introduction to the control aspect of platoon maneuvers in Automated Highway Systems. The different platoon maneuvers include lane changes, merge procedures and split procedures. The first part of the report consists of a review of the existing literature in this area. The survey is split up into the quasi-synchronous and vehicle follower controller methods of approach to the problem. The subsequent sections deal with the mathematical description of the problem with two representations of the system under study, varying only in level of model complexity. The simplified model has been studied with a linear based control as well as the nonlinear sliding mode approach. The results only reinforce the need for nonlinear systems analysis and control. The sliding control technique has also been applied to study the detailed version of the plant.