WiNEST provides new and unique opportunities for experimentation across client devices, edge devices and in-network compute resources in a large, city-wide geographical area typical to a mobile network operator or wireless ISP. In cooperation with our industry partner, 5Nines Data, we expect to cover a large area across the city of Madison.
WiNest is focused on building infrastructure to provide computing resources in the Madison area. In the above map, this displays the current predicted 4G coverage with our current deployments. The management of the base stations is done in conjunction with 5nines LLC. Using the 4G coverage, we are able to deploy fixed ParaDrop compute resources around the city. These compute resources are highly flexible and programmable for supporting research in many disciplines. With fixed resources located around State Street and the capitol areas, there are opportunities experimentation in a real world scenario. Further deployments currently underway are wireless resources and compute resources along Park Street.
To provide high speed and low latency to each of the computing resources, we are deploying LTE to South Park St. The map below identifies potential locations identified where power and network connectivity are present in the city. We are planning to install 25-30 LTE pico cells in total for full low-latency, high bandwidth up and down the street.
We intend to have wireless resources in research vehicles along this route to test various driving scenarios. With ParaDrop in the Innova Urban Electric Vehicles, we will have hotspots for the occupants to be able to test applications in mobile environments with local computing resources located inside the vehicle. In addition to the UEV, ParaDrop resources will be located in Madison Metro city buses throughout the city.
Here we will cover the basic idea of how to access ParaDrop resources.
We developed a low cost Wi-Fi-based in vehicle monitoring and tracking system that can passively observe mobile devices and provide various analytics about people both within and outside a vehicle which can lead to interesting population insights at a city scale. Our system leverages WiNEST architecture and runs on a vehicle-based edge computing platform. The vehicle-based edge computing platform provides computing and storage resources which allow us to process massive amount of data in a timely manner. Trellis allows operators to collect various information and conduct spatial-temporal analytics in real time. With the help of WiNEST and ParaDrop, various relevant transit analytics, such as popular origin-destination stations, occupancy of vehicles, pedestrian activity trends, can be quickly derived on-board and sent back to transit operators without incurring high data requirements from the vehicles. What’s more, it is easy to deploy and manage such applications in multiple vehicles across a whole city using WiNEST and ParaDrop.
We deploy Trellis in two city buses. Those two buses have been assigned to three bus routes that are illustrated in figures below. The bus routes cover the main campus area of University of Wisconsin Madison, as well as a residential area that accommodates graduate students and visiting scholars. These two buses are usually scheduled to be on the road from 6am to 6pm on route 80. Buses are also occasionally scheduled to operate during night hours on route 81 and 82. We collected data from both buses for around 300 days for 12 hours per day. In total, during these 300 days, two buses travel more than 32,000 miles. Among the collected data traces, the two buses ran on route 80, 81 and 82 for 258, 23, and 24 days accordingly. More than 300,000 unique Wi-Fi devices were detected by our system.