Privacy Preservation Research Based On Pseudonym Changing In VANETs

The Vehicular Ad-hoc Networks(VANETs)are able to connect people,vehicles,roads and other traffic factors through wireless communication.It can effectively enhance safe driving,optimize transportation efficiency and improve driving experience.In order to achieve safety-related driving applications,vehicles need to periodically broadcast beacons containing their driving status.By receiving beacons and performing dynamic processing,the surrounding vehicles can sense or predict their traffic conditions in real time.However,typical data encryption mechanism is not applied to beacons in communication process Although the real-time data transmission is guaranteed,beacons can be exploited easily by an attacker and utilized to infer the users' privacy.As a privacy protection scheme under such problem,pseudonym changing uses a pseudonym to replace the vehicle's real identity.Periodically changing can not only enhance the non-relevance between a single pseudonym and the identity,but also ensure the user's location privacy.However,there is currently no general evaluation framework to evaluate the pseudonym changing methods,and compare the experimental results between different strategies.In view of this,this thesis designs an evaluation scheme for pseudonym changing by setting up a simulation experiment platform for VANETs,and carries out evaluation testing and further improvement research on pseudonym changing methodsThe main contents of the thesis are as following(1)We first design and build a simulation-based platform for privacy research in VANETs.The platform is composed of five main parts:traffic simulator,communication module,privacy preservation algorithm and evaluation module,visual interface and traffic data.The platform has comprehensive functions,strong expandability and easy experiment operation,which provides a basis for the engineering implementation and evaluation tests of the pseudonym changing methods(2)Based on the simulation experiment platform,we then carry out an experimental study on typical pseudonym changing methods.First,the existing methods are classified according to the conditions of pseudonym changing.According to the characteristics and processes of the pseudonym changing in each type,the programs of seven algorithms are implemented by using Python.Finally,the experimental effect of the pseudonym changing is shown,and the experimental data used for subsequent evaluation is also exported(3)In order to carry out the evaluation of pseudonym changing methods,we design a general evaluation scheme.The scheme includes several main parts such as adversary model,data preprocessing,state prediction,etc.The state information in beacons is used for prediction,thereby achieving the linking between different pseudonyms.In addition,the evaluation plan also introduces four different types of side information,and analyzes the impact on the results from the theoretical and experimental perspectives.