Research On Task Participant Selection Menthod In Mobile Crowd Sensing

Nowadays,the speed of mobile terminal equipment upgrades is getting faster and faster.The types of mobile terminal devices have changed from the original single mode to the current innumerable flowers.There are countless possibilities.This is due to the rapid development of integrated circuit technology.Can talk easily to now can handle a large amount of data information in a short time.In recent years,a perception mode called Mobile Crowd Sensing(MCS)has gradually appeared in major academic journal articles.This mode is very similar to the well-known IoT perception mode in terms of components and workflow.At the same time,it is also a perception mode combining mobile sensing technology and crowdsourcing technology.With scholars' research on mobile crowd sensing perception technology,many platforms or organizations also see its rich use value,and have organized teams to study its application systems in daily life.Because the sensing device of mobile crowd sensing technology is an intelligent mobile device or vehicle-mounted device that people will use in their lives,as long as anyone who owns these devices can sign up to participate in the perception task to complete a specified operation and get a certain reward.In order to complete a certain sensing task and obtain the desired data,it is necessary to control the choice of participants,because the platform or organization has different application areas,so the characteristics of the participants are not the same.Some participants need to be able to provide high-quality perception data,so they will pay attention to their historical perception data;some participants who need high response speed,will value the time of their historical receiving perception tasks.The main content of this article is to propose a more suitable method for selecting participants based on the characteristics of different perception tasks.The main contents include the following:A method for selecting participants based on a trust policy is proposed.First,based on the trust policy proposed by the user,attribute reduction is used to delete the participant's trust attributes that are not important to the perceived task,and simplify the multi-attribute index that affects the participants' trust degree.Then the gray trust analysis method is used to compare the selected trust attribute with the target trust attribute to obtain the participant's current trust degree.Finally,a dynamic trustassessment model for participants in a complex network environment(DTEM-GCA for short)is established.The simulation test results show that the proposed trust evaluation model can not only draw the participants' trust in real time,but also have higher accuracy and less time.A data quality-based participant selection method(herein called PSM-DQ)was proposed.The fuzzy inference method was used to classify the perceptual tasks into different levels according to the data quality requirements of the perceptual tasks.Then the participants' trust evaluation criteria were established.The data quality is combined with the time decay factor to calculate the current data quality,and then the historical trust is used to update the participant's current trust degree.Finally,the punishment mechanism is used to reduce the trust of the malicious participant and discard it,thereby ensuring the perceived data quality to a certain extent.Simulation experiments prove that PSM-DQ can recruit more participants and improve the perceived data quality to a certain extent.