Crowdsensing From the Perspective of Behavioral Economics: An Incentive Mechanism Based on Mental Accounting in IoT
Crowdsensing is a new paradigm of applications that takes advantage of mobile devices to collect sensing data. Tasks in crowdsensing will consume users' resources so that incentive mechanisms are necessary to encourage users' participation. Existing incentive mechanisms are based on traditional economics, which have two common problems: 1) the utility of different tasks is fungible and 2) users' behavioral preferences are consistent. Mental accounting (MA) theory in behavioral economics proves that the utility of tasks obtained in different ways is nonfungible and people's preferences of behavior are inconsistent. Reference dependence, loss aversion, and sensitivity decline are the three characteristics of MA. Reference dependence means people evaluate outcomes relative to a reference point, and then classify gains and losses. Loss aversion refers to people's tendency to prefer avoiding losses to acquiring equivalent gains. Sensitivity decline means that the marginal utility of MA about gains and losses is diminishing. Thus, this paper proposes an incentive mechanism called the MA auction incentive mechanism (MAAIM). Based on reference dependence, coupled with sensitivity decline, we establish an external reference environment and an internal reference point to motivate users. Based on loss aversion, we design a payment mechanism to encourage users to improve their data quality. The extensive simulation results show that MAAIM improves the number of users participating in sensing tasks, the utility of the sensing platform, and the quality of data collected by users.