Date of Award:
Master of Science (MS)
Electrical and Computer Engineering
Vehicle platooning has been heavily studied the last decade. A transportation system formed by electric vehicles driven by control systems with the help of on-board sensors, wireless inter-vehicle communication, and wireless recharge capability has been shown to increase highway capacity, transportation safety, reduce travel time, save energy, and release human drivers from stress. Two layers of control are required to automate a platoon, the low-level vehicle control, and the upper-level platoon control which seeks to maintain the constant spacing of the platoon, and avoid collisions. In order to have a robust platoon, the vehicle control system needs to be robust to gain variations. Simulations were run in Matlab's Simulink to compare how well a vehicle control system would behave in the presences of nonlinearities and disturbances. The integer order and fractional order controllers were designed with the same specications. Fractional order controllers present better performance with no overshoot for the speed servo, and faster response for the steering system. For platoon control, the necessity is to achieve string stability. The bi-directional and leader-follower architectures have been shown to achieve string stability. Still, what happens to all the benets of platooning when a malicious vehicle (attacker) attempts to perturb the system? This malicious attack could be the result of a company trying to sabotage the operation of another's in order to make it spend more energy than required, and thus raise its transportation costs. By using Matlab, a simulation platform was designed. It was used to simulate the response of a robust platoon to an optimal attack prole, generated by Matlab's genetic algorithm. To calculate the energy expenditure a model for a 1995 Honda Accord LX from cappielo's analysis is used. Two scenarios are considered: 1) the attacker intends to make the whole platoon spend extra energy, and 2) the attacker focuses on aecting only one victim. The greatest amount of extra energy expenditure for the rst scenario was obtained with the bi-directional architecture and a size 3 platoon (140%). The leader-follower architecture limited this peak value to 94% for a size 8 platoon. In order to really prot from the benets of platooning, a platoon size 8 or more is recommended. In this desirable range, the bi-directional control law manages to limit the extra energy expenditure to 80% (size 8) to only 35% (size 20). For the leader-follower and a size 20 platoon, the optimal attack produced an extra 65% expenditure. For the second scenario, with the bi-directional architecture the attacker could make the victim spend up to 122% (size 10). Still, this depends on both the attacker's and the victim's position. For instance, with the attacker in position 2, only 8% extra energy was observed. The leader-follower architecture allowed between 80% to 110% in any position for the attacker while in front of the victim (the attacker cannot aect the victim from behind). Regenerative braking in all cases saved between 35% to 50% of the energy that would be otherwise lost by the use of dissipative brakes. In order to create an operational platoon system, that is as robust as possible to the attack, the recommended platoon size is 12 or more. The use of regenerative braking capable vehicles is a must. The control system should be the fastest possible, and make use of the bi-directional architecture to limit energy expenditure. The implementation of an attacker or defective vehicle detection system is recommend, taking the measure of making the attacker=defective vehicle reposition to the last in the platoon.
Cornelio Sosa, David A., "An Efficiency-Motivated Attack Against Vehicles in a Platoon: Local Vehicle Control, Platoon Control Strategies, and Drive Train Technologies Considerations" (2014). All Graduate Theses and Dissertations. Paper 2168.
Copyright for this work is retained by the student. If you have any questions regarding the inclusion of this work in the Digital Commons, please email us at .