Pervasive Mobile Video Alarm & Surveillance Network and System

Description

Atragedyhappened on Oct. 13th, 2011, in FoShan City, Guangdong Province, mainland China. Little Yueyue, “the toddler who made global headlines after she was hit by two cars and left unassisted as more than a dozen pedestrians passed her by, has died”*. A surveillance camera recorded the cars and subsequent pedestrians. Had this camera sent a warning video to the drivers’ mobile devices beforehand or pushed an alarm video to nearby police’s mobile devices, the accident might result in a different result (or Little Yueyue could have been saved)! Apart from traffic surveillance, such pervasive (ubiquitous) mobile Video Alarm & Surveillance (VAS) services should be widely useful for home, business, public society security and hash environment surveillances such as wildlife parks, forests, country board safeguard or military actions etc. Unfortunately, today’s surveillance systems (CCTV) are not advanced to this step. Current wireless VAS researches mainly concern with bandwidth and power usages. In our opinion, jointly efforts of fast push/pull of VAS flows to/from various end-devices, traversing through heterogeneous wired/wireless networks with limited bandwidth, while the flows have to be stored on the fly, are more critical that have been largely ignored.To fill-in the gaps, in this project, we intend to design and prototype a novel Pervasive Mobile Video Alarm & Surveillance System (PMVASS) to provide VAS services for end users particularly with mobile devices. PMVASS will be materialized through associating a set of Pervasive Objects (POs), equipped with high-definition (HD) visual camera and multiple interfaces (3G/4G, WiFi or Ethernet) and can be deployed in unplanned manner anywhere and collaborate in wired/wireless networks to provide mobile VAS services. We will carry on the composition research of efficient VAS flow service and storage constrained by heterogeneous networks/devices with limited bandwidth. Specific research tasks include: (1) Novel Pervasive Object design, association and optimal admission control algorithms will be sought to admit the mobile user devices for VAS push/pull services, taking PO’s available resources (i.e., bandwidth) and nature of applications into considerations. We will prioritize the admission controls for mobile users’ who already registered the VAS services and explore the best “matching” admissions for POs and mobile users; (2) Novel anycast algorithms enable POs and mobile devices to transmit (via push/pull) the VAS flow to/from a group of (replicated) Networked Video Recorder (NVR) servers. Mobile devices assisted offloading mechanism will be sought using the NVR anycast group address; (3) Distributed anycast algorithms with maximum flow analysis and theoretical bounds will be investigated which differs fundamentally from the traditional maximum flow problem in the sense that the flows can be maximized through exploring multiple anycast paths; (4) Prototypes of PO firmware, terminal software (Smartphones or normal videophones etc.) and protocols will be implemented based on existing video coding, tracking and recognition techniques to acquire real experimental data for demonstrating the feasibility and performance of PMVASS.Completion of this project will help Hong Kong to an irreplaceable position in the novel concept, algorithm designs, system development and knowledge transfer in advancing the core technology of mobile VAS services for large number of critical applications. With the expected outcomes, the project will bring HK with benefits of 1) Novel concepts, algorithms, protocol design, platform and systems to deal with the related fundamental issues in wide spectrum; 2) High quality publications and Intellectual Property generated from the project may also help HK's industries to upgrade their high-tech level world-wide.* http://blogs.wsj.com/chinarealtime/2011/10/21/little-yueyue-dies-amid-soul-searching/