Context-Aware Mobile Crowdsourcing and Logistics

Abstract/Technology Overview

Mobile crowdsourcing continues to suffer from problems such as unpredictable and unfair task completion rates and worker churn. We believe that such deficiencies can only be rectified via a centrally-coordinated crowdsourcing approach, where tasks are recommended (pushed) to people based on their predicted movement trajectories. Moreover, our crowdsourcing technologies allow tasks to be aggregated into bundles (which amortize the travel overhead across multiple tasks) and support several effective forms of variable task pricing. These technologies can effectively increase both the task acceptance and completion rate for a variety of location-centric crowdsourcing platforms, and thereby help to significantly increase the participation rate of residents in tackling urban operations challenges.

Technology Features, Specifications and Advantages

Trajectory aware proactive task recommendation: Using the historical traces of individual user movement, a predictive trajectory profile is developed. Based on the prediction, a set of tasks that are best suited to each individual’s predicted movement pattern is generated, so that the extra detour incurred is kept to the minimum.

Task Bundling and Variable Pricing: Uses the history of worker movement and task completion behaviour to create different types of task bundles and automatically adjust the reward prices for tasks, so as to ensure that tasks are completed more uniformly across the entire geographical area.

Dynamic Task Offloading: Allows community-supported crowdsourcing practices, by enabling workers to dynamically offload previously accepted tasks to willing buddies and share the resulting task completion rewards.

Potential Applications

This technology is applicable in the following areas:

Retail sensing

  • Crowd-sourced queue and inventory monitoring
  • Store display compliance
  • Last-mile crowdsourced package delivery

Urban operations and monitoring

  • Municipal services monitoring
  • Fleet diagnostics
  • Bike path monitoring (usage patterns of trails, bumps/blocks on trails)
  • Smart campus monitoring in universities, offices, housing estates, etc

Customer Benefit

  • The system also helps us to target the partial development of a smart campus, where a volunteer student population is used to perform continuous sensing of campus resources. This can be closely translated to a city-scale monitoring.

  • Customers can utilize the various innovative and automated features of the platform to significantly increase the participation rate of crowdworkers, as well as their eventual task completion rate.

  • Our technologies have been proven to improve task worker productivity—letting them earn more money while incurring much lower (50% less) travel detour overhead.

Contact Person

Janice Tan


Singapore Management University - Living Analytics Research Centre

Technology Category
  • Ambient Intelligence & Context-Aware Computing
  • Social Media, Collaboration & Crowdsourcing
Technology Readiness Level
  • TRL 7