Institute of Bioorganic Chemistry of the Polish Academy of Sciences, Poznań Supercomputing and Networking Center (PSNC)


Cloud Eyes is a Smart City Security Monitoring Platform enabling real-time video surveillance using WiFi and mobile broadband (MBB) networks, with dynamic adaptation of transmission parameters. It is designed to provide monitoring from buses, trams, trains, robots and places where connection using a wired network is not possible. The positive feedback from potential customers made us decide to create a market ready product. However before going to the market, we needed to perform validation of the platform in a real-world MBB environment. That is why we had chosen to use the MONROE testbed to deploy Cloud Eyes in a distributed system of mobile and stationary nodes connected via various wireless networks, and to measure the key network metrics and the corresponding video quality. Our goal has been to determine the correlation between Quality of Service and Quality of Experience, define encoding and transmission profiles for various network conditions and enhance our smart adaptive algorithms, making Cloud Eyes a market-ready solution running on heterogeneous wireless networks.

During the project we ran a series of experiments using stationary and mobile nodes located in Sweden, Norway and Italy. Each experiment used a pre-encoded series of video sequences, uploaded to a MONROE node within the experiment Docker image. The sequences were streamed using mobile networks to our security monitoring center in Poznań. In the center, our automatic Quality of Experience tool measured the quality of the video. We correlated this with the measured Quality of Service parameters and tuned the encoding and transmission profiles for each set of values. The Quality of Experience tool uses advanced no-reference metrics to evaluate the quality of digital images without access to non-distorted reference images or any features representing the reference images. Instead of making comparisons, no-reference techniques try to identify or predict impairments such as blockiness or jerkiness by analysing the characteristics of the underlying network transport and by analysing the video itself. The videos were also streamed to the monitoring centre, where security monitoring personnel conducted subjective expert assessment of the video quality and its usefulness. Furthermore, we bought a MONROE node to conduct tests involving Polish mobile network operators. We equipped our node with SIM cards from different operators and conducted tests with the MONROE node installed in a vehicle. This helped us assess performance in Poland, our initial target market.

Our product is beyond the state-of-the-art of currently available monitoring solutions as it allows for continuous, real-time surveillance, utilising both WiFi and MBB networks, whereas our competitors’ solutions rely only on WiFi and are not capable of continuous, real-time monitoring. Our product takes advantage of the full capacity of MBB and overcomes problems of varying transmission quality by using dynamic adaptation. During the project we have presented the Cloud Eyes platform and the results of the experimentation at fairs and conferences related to security monitoring and ICT. We also installed Cloud Eyes in our living-lab environment and presented the platform to representatives of public transportation companies, City of Poznań authorities and companies interested in innovative surveillance systems. Based on the feedback from end-users and potential customers, we have prepared an outline of our business model.

The experimentation conducted within the MONROE testbed allowed us to tune the encoding and transmission profiles, as well as maximize the quality and usefulness of surveillance video in changing mobile network conditions. Furthermore, the experiments helped us calibrate the automatic Quality of Experience measurement tool. Finally, we have verified that our technology can provide continuous, real-time surveillance from public transportation vehicles, overcoming problems related to video streaming over mobile networks. The incorporation of these results in the Cloud Eyes platform will allow us to accelerate the process of preparing a market ready solution.