Research On Task Assignment And Security Mechanism In Mobile Crowdsensing System

Mobile crowdsensing is a paradigm in which mobile users can cooperatively col-lect data and complete complex tasks using their portable electronic device.In recent years,with the popularity of smart phones,on-board devices,and wearable devices,a large amount of important information can be collected.Since mobile devices can be utilized to conduct sensing tasks and we can use idle resources to collect data.It has s-timulated many applications,,such as noise mapping,environment mapping,traffic flow mapping,emergency event reminding,book digitization,and so on.Therefore it has a great commercial value and research significance.Most of the existing theoretical researches focus on the privacy preserving mech-anism of sensing data and location information,and there is no research on the privacy preserving mechanism of users involved in the task allocation process.In addition,there are few researches on user authentication mechanisms for illegal users participating in the system.Therefore,in this dissertation,we investigate the task allocation problem and related security mechanisms in mobile crowdsensing systems,including privacy p-reserving and user authentication.First,we see time duration as a kind of sensing quality and study the time sensitive task allocation mechanism in mobile crowdsensing sys-tems,and investigate the corresponding optimization mechanism;Second,we study the security mechanisms of mobile crowdsensing system,that is,the privacy preserving mechanism of users involved in the task allocation process and the user authentication mechanism in mobile crowdsensing system.Specifically,the main contributions to this dissertation are summarized as follows:1.We propose a time sensitive task allocation mechanism for mobile crowdsensing systems.We prove that the task allocation problem is NP-hard.Then,we design a greedy algorithm to solve the problem,and further have optimized the solution by using tabu search.Experiment shows that this proposed task allocation mechanism can reduce the cost by about 10%in average while recruiting users to perform tasks compared with the existing greedy algorithms.2.We propose a privacy preserving task allocation protocol.We form the sensing quality sensitive task into a weighted set coverage problem and design a basic task allocation algorithm based on greedy strategy.Then,the task allocation is regarded as a secure multi-party computation process.We use the secret sharing technique to design a task allocation protocol under the semi honest model.Finally,the security of the protocol is proved,and the approximate ratio of its task allocation performance is also analyzed.Compared with the existing works,the protocol does need any trusted third party,and any the encryption or decryption operations,so it has better computing performance.3.We propose a user authentication mechanism based on gesture detection for mobile crowdsensing systems.We analyze the existing authentication mechanisms,and use spectrogram to present the Doppler effect of the gesture based on the sonic wave.Then we use target detection technology to authorize users.Our experiment shows that this method has reliable the accuracy of ultrasonic gesture recognition,and gives an acceptable authentication mechanism.