Crowdsensing in the Wild with Aliens and Micro-payments Manoop Talasila, Reza Curtmola, and Cristian Borcea Computer Science Department New Jersey Institute of Technology Newark, NJ, USA Email: mt57@njit.edu, crix@njit.edu, borcea@njit.edu Abstract—This article presents results and lessons learned from two user studies on crowdsensing incentives, specifically mobile gaming and micro-payments. The analysis of the results suggests that gaming is a cost-effective solution for uniform area coverage, while micro-payments work well for sensing tasks with tight time constraints or long-term tasks for personal analytics. I. I NTRODUCTION Mobile crowdsensing can be used to enable a broad spec- trum of applications, ranging from monitoring pollution or traffic in cities to epidemic disease monitoring or reporting from disaster situations. Several mobile apps [1] and plat- forms [2], [3] have recently been proposed for crowdsensing. A major challenge for broader adoption of mobile crowd- sensing apps/systems is how to incentivize people to collect and share sensor data. In addition, uniform sensing across a target area is a desirable property in many cases, and incentives need to be provided to collect data from unpopular regions of that area. Many of the proposed mobile crowdsensing systems provide monetary incentives to smart phone users to collect sensing data. There are solutions based on micro-payments [4] in which small tasks are matched with small payments. Other techniques were also explored to motivate individuals to partic- ipate in sensing. For example, beneficial personal analytics are provided as incentives to participants through sharing bicycle ride details in Biketastic [5]. Another variety of incentive is enabling data bartering to obtain additional information (bargain hunting through price queries in LiveCompare [6]). In addition, there are gamification techniques proposed for crowdsourced applications [7]. We recently designed, built, and evaluated through user studies two crowdsensing systems based on different types of incentives: (1) a mobile game that uses in-game incentives to convince participants to cover all the regions of a target area, including the unpopular ones; (2) a micro-payment based system that allows users to pick the sensing tasks they want to execute according to their own criteria and provides a small payment for each task. Our game, “Alien vs. Mobile User” [8], is a first person shooter sensing game which is played by mobile crowdsensing participants on their smart phones. The game involves tracking the location of extraterrestrial aliens on the campus map of our institution and destroying them. The game entices users to un- popular regions through a combination of in-game incentives, which include alien-finding hints and higher number of points received for destroying aliens in these regions. The game was implemented in Android, and it collects WiFi signal data to construct the WiFi coverage map of the targeted area. Our micro-payment based system, McSense [3], allows the participants to choose from a wide-range of sensing tasks such as taking photos at events on campus, collecting GPS and accelerometer readings, or collecting application and network usage from the phones. When choosing a task, the participants have to balance the value of micro-payment (different for each task) against their effort, the potential loss in privacy and the resource consumption on the phone (e.g., battery). A McSense application was implemented for Android phones. This article describes the results and lessons learned from two user studies, one for gaming and one for micro-payments. Any student on campus was allowed to participate in these studies, and all they had to do in order to participate was to install our mobile apps and sign a participant form. Over 50 students participated in each study, and the duration of the studies were 35 days for mobile gaming and two months for micro-payments. II. RELATED WORK Micro-payments have been used as an incentive for users to complete tasks in crowdsourcing (e.g., Amazon MTurk - http://www.mturk.com). Micro-payments have also been ex- amined in the context of participatory sensing [4] and crowd- searching [9]. The work in [10] presents pricing incentive mechanisms to collect quality data in participatory sensing applications. Some of the key findings are that incentives can be highly beneficial in recruiting participants and that micro- payments have the potential to extend participant coverage both spatially and temporally. In addition to these insights, our micro-payment based study identifies data reliability as a significant issue in crowdsensing based on micro-payments. Attracting people to unpopular places could be difficult and expensive. For example, the results from a recent crowdsensing study [11] show that many places will be infrequently visited. Gaming could be a cost effective alternative to micro-payments when attracting users to unpopular regions is necessary. In the same context, the concept of steered crowdsensing is proposed in [12] to address the data quality issue in crowd- sensing. The paper argues that most crowdsensing systems are 1