Context Optimized recruitment and predictive analytics for MCS recruitment strategy.
The emergence of sensor-rich smartphones has expanded the growth of Mobile Crowd Sensing (MCS). Presence of the required context is vital for mobile crowd sensing. This research focuses on the development of a distributed architecture for executing mobile crowd sensing applications in real-time. Emerging MCS applications such as tracking a perpetrator in real time, surveillance and monitoring disaster or an event have analytical components that can be processed on edge devices as the event is in progress.
The ContextAiDe architecture is a middleware that enables easy development of real-time MCS applications. Important directions for the design of ContextAiDe are: a. Model the context based on requirements of this applications and matching context during recruitment b. Optimization of operational overheads involved in complex analytical processing tasks executed on edge devices, c. Proactive strategies to reduce task failures (Details of Data Analysis and Prediction used in various apps supported by ContextAiDe can be found here).
ContextAiDe middleware is demonstrated using a Perpetrator Tracking app. This app is designed using ContextAiDe to track a perpetrator as he moves from one location to another. The idea is to acquire phone/video from devices near to tracking location and update the location as the perpetrator moves. Task such as image capture, face detection are executed on the mobile device while the face recognition task are run on Edge/Cloud Server.
- Madhurima Pore, Vinaya Chakati, Ayan Banerjee, Sandeep K.S. Gupta. ContextAiDe: End to End Architecture for Mobile Crowd Sensing Applications. ACM Trans. Internet Technol. 9, 4, Article 39 (December 2017), 20 pages (Accepted for publication).
- Pore, M., Chakati, V., Banerjee, A., & Gupta, S. K. (2016, October). ContextMete: Context Optimized Peer Offload Architecture for Distributed Mobile Applications. In 2016 IEEE/ACM Symposium on Edge Computing (SEC) (pp. 105-106). IEEE. (pdf)
- Pore, M., Sadeghi, K., Chakati, V., Banerjee, A., & Gupta, S. K. (2015, September). Enabling real-time collaborative brain-mobile interactive applications on volunteer mobile devices. In Proceedings of the 2nd International Workshop on Hot Topics in Wireless (pp. 46-50). ACM. (pdf)
- Madhurima Pore, Vinaya Chakati, Ayan Banerjee, Sandeep K.S. Gupta, Middleware for Fog and Edge Computing: Design Issues, Wiley STM: Fog and Edge Computing: Principles and Paradigms, January 2018 (accepted for publication).
Researchers: Madhurima Pore, Vinaya Chakati, Dr. Ayan Banerjee, Dr. Sandeep K.S. Gupta