Crowdsourcing for On-street Smart Parking Xiao Chen, Elizeu Santos-Neto, Matei Ripeanu Department of Electrical and Computer Engineering, University of British Columbia {xiaoc, elizeus,matei}@ece.ubc.ca ABSTRACT Crowdsourcing has inspired a variety of novel mobile applications. However, identifying common practices across different applications is still challenging. In this paper, we use smart parking as a case study to investigate features of crowdsourcing that may apply to other mobile applications. Based on this we derive principles for efficiently harnessing crowdsourcing. We draw three key guidelines: First, we suggest that that the organizer can play an important role in coordinating participants’, a key factor to successful crowdsourcing experience. Second, we suggest that the expected participation rate is a key factor when designing the crowdsourcing system: a system with a lower expected participation rate will place a higher burden in individual participants (e.g., through more complex interfaces that aim to improve the accuracy of the collected data). Finally, we suggest that not only above certain threshold of contributors, a crowdsourcing-based system is resilient to freeriding but, surprisingly, that including freeriders (i.e., actors that do not participate in system effort but share its benefits in terms of coordination) benefits the entire system. Categories and Subject Descriptors C.2.4 [Computer Systems Organization]: COMPUTER- COMMUNICATION NETWORKS - Distributed Systems - Distributed applications General Terms Algorithms, Design, Economics, Human Factors Keywords Mobile crowdsourcing, smart parking, collaborative sensing. 1. INTRODUCTION In many application contexts, crowdsourcing has reintroduced humans into the information processing loop. Several success stories [1,2] show that when ordinary people’s knowledge, experience, or creativity are aggregated, they have the potential to replace the roles of the most powerful computers, the most knowledgeable experts, or the most talented designers. Currently, the majority of crowdsourcing-based applications focus on harnessing collective intelligence via web applications. However, as the wireless communication infrastructure and mobile applications keep thriving these years, the influence of crowdsourcing can have direct impact on applications affect our physical world as well. When equipped with a smartphone and constantly connected to the wireless network, everyone is able to collect real-time data about the physical world either through her observation and manual input or by the sensors in the device. Therefore, mobile crowdsourcing enables data collection through thousands or millions of such intelligent probes and collects data primarily from the surroundings of people’s everyday life. Such collaborative data collection enables the design and implementation of services that are helpful to each individual in our increasingly connected society. One example is the realization of smart parking. The parking problem has existed in big cities around the world for decades. Studies show that an average of 30% of the traffic [3] in busy areas is caused by vehicles cruising for vacant parking spots. The situation is getting even worse in developing countries like China, where the number of private cars soars without sufficient parking facilities. The extra traffic causes a series of problems such as traffic congestions and accidents, air pollution and waste of fuel. Some local governments try to mitigate these issues by means of smart parking. In a nutshell, they try to employ information and communication technologies to collect and distribute the real-time data about parking availability so that drivers can spot the right parking place quickly. For example, the city of San Francisco installed thousands of sensors to on-street parking spaces in its busy areas for parking management. While the effect of such a centralized approach is immediate, its huge initial investment and maintenance cost inhibits a widespread adoption in most other cities. In this paper, we derive design guidelines for a mobile crowdsourcing framework that integrates the functionality of parking guidance into a road navigation system. To be able to generalize our finding to other crowd-sensing scenarios, we assume that there is no additional sensing infrastructure deployed to monitor parking spaces. The only sensing device our solution requires is the road navigation system each driver alreay uses. While we are not the first to propose crowdsourcing for smart parking, our proposal has several advantages over existing approaches. First, by crowdsourcing through a road navigation system, we eliminate unnecessary manual operation during driving, which complies with safety regulation in most countries. Unlike applications such as Open Spot [8], which requires drivers to launch it separately for searching parking spots, we only ask drivers for their manual input at the beginning and the end of their trips. By simplifying operations, we are more likely to recruit a larger number of contributors, a key success factor for crowdsourcing. Second, since drivers who contribute can also benefit from the system, our approach heavily depends on a pattern of mutual assistance, which excludes the complexities caused by monetary reward [5]. On the one hand, the system has a high resilience to the existence of free riders as shown in the following sections. On the other hand, because we assume a centralized control over the data distribution, we can differentiate users with different quality of service based on their contribution records. It could serve as a supplement mechanism to enforce fairness and motivate participants to contribute. Third, we coordinate the crowdsourcing behavior among participants to boost the efficiency of the data collection and utilization. In contrast to existing approaches that only share information about parking vacancies, our system also tries to