article

Understanding the Science Behind Small Cell Deployment

by: Sarat Puthenpura, November 7, 2013
Small-cell-graphic

 

The continued surge in mobile data demand, paired with the finite nature of spectrum resources, and the desire to provide customers with the highest quality service has landed network operators in an interesting position over the last few years. In response to these needs, several carriers have turned to small cell technology as one solution to address these challenges. Small cells are low-power radio access nodes that provide a resourceful network solution for coverage, capacity, and quality.

It’s been commonly recognized for several years that small cells are going to make a big impact on mobile networks, as they’re able to increase spatial reuse of existing licensed spectrum. While the concept isn’t new, the necessity of this technology has become apparent with growth in mobile data traffic increasing by more than 30,000 percent on AT&T’s network from January 2007 through December 2012. Small cells are quickly becoming a flexible and essential component of our network. As we continue to integrate this technology into our living, breathing network we’ve seen an improvement in network coverage and capacity.

201311_Small_Cells_Fig_01_Outdoor

As carriers are working to enhance the network, we’re seeing infrastructure move closer to customers. And small cells are no exception. Traditional, or macro cells, sit on top of the landscape and send and receive signals from above the buildings, trees, bridges and more. Small cells, on the other hand, are placed within the “clutter” of the landscape, and signals to and from them travel through obstructions – like trees and buildings.  Hence the radio frequency (RF) propagation characteristics of small cells are very different from that of traditional macro cells. While small cells are not a replacement for macro cells, or answer to current spectrum demand, they’ve quickly become an integral part of the network.

However, with this technology comes a whole new set of network planning considerations – such as understanding the radio frequency (RF) propagation for small cells. While the models for deployment of traditional macro cells are well studied, insight into placement of small cells – both indoor and outdoor – has been relatively unknown until now.

The Charge

Seeing that the future of the network includes a small cell layer and wanting to ensure that customers receive the maximum benefit from this technology, researchers realized that there was a lot of work to be done before the network planning and implementation could begin. Scientists and engineers at AT&T Labs were charged with understanding the science behind the deployment of this technology.

The team, knowing that current network planning models for macro cells wouldn’t work for small cells, embarked on a task to identify RF propagation models based on which small cell planning tools can be built. In order to accomplish this, they needed to understand the characteristics of RF propagation for small cells, both indoor and outdoor, in the real world. This knowledge would help to ensure that the technology is positioned correctly in the first place, providing enhanced service quality to our customers.

201311_Small_Cells_Fig_02_Plan

Following their objective for creating a planning capability for small cells, researchers conducted experiments separately for both indoor and outdoor to figure out appropriate RF models. A crew of five researchers, including Executive Director at AT&T Labs Sarat Puthenpura, as well as Vaneet Aggarwal, Rittwik Jana, N.K. Shankaranarayanan, and Ioannis Broustis, formulated a plan involving a van with a retractable mast,  that would allow them to test small cell propagation outdoors at various antenna heights. Similarly for indoors, they created a testing platform housed on a mobile rack so that they could move the small cell setup around and assess the RF propagation characteristics in real office environments.

In both cases, the team had to take hundreds of measurements in the coverage area of small cells. They accomplished by moving around the equipment, which continuously logged key RF measurements, while keeping track of locations.

The researchers conducted their first experiment in Morristown, NJ, nearby their office location in Florham Park, NJ.  Prior to doing outdoor experiments, the researchers informed local police about their plans and got their help and advice. “We were worried that we would have neighborhoods concerned about a truck with a mast going up and people running around with laptops,” said Puthenpura.  In order to be sensitive to local people and traffic, the researchers conducted their experiments early in the morning before rush hours.

Subsequently, the team then did outdoor experiments in downtown Manhattan and later collected extensive measurements from several major cities with the help of a drive test vendor.

201311_Small_Cells_Fig_03_transmitter

The team used digital terrain data, including building heights and locations, to develop a model which was used to create a tool to calculate the number of small cells needed to optimally cover a particular area.

201311_Small_Cells_Fig_04_indoor

The team then moved to look at propagation models for small cells deployed in indoor locations.  They used their Florham Park office for this purpose. Instead of driving a truck around, in this case they had to push a small cell setup on a mobile rack.  One good thing here was, there was no need to contact the police department!

To analyze the data and create models, instead of digital terrain maps, the team used digital building floor plans. 

 

“We went into the street, did the testing, and we know how small cells work – that’s a confidence our customer’s should have,”

 The Outcome

After concluding their tests, both indoors and outdoors, researchers saw that the fundamentals of both models were relatively similar. Indoors the signal had to go through and around walls, cubicles and various other barriers; while outdoors there are similar obstructions, just in the forms of buildings, corners and trees. This was especially good for planning tool development since the same model could be used for planning both indoor and outdoor small cells.

As researchers had assumed, the radio propagation models for small cells were vastly different from that of macro cells. Through studying radio propagation and mathematical modeling, they could come up with a single model that would enable them to place small cells balancing the three fundamental aspects of cell planning – capacity, coverage and quality. This led them to support the development of AT&T’s small cell planning tool, the Hetnet Analysis and Resource Planning – or HARP – tool.  HARP allows network planners to input the traffic demands, along with building floor plans or topography maps, and see how the expected coverage is for both indoor and outdoor. Thus the tool can tell planners the optimal number and placement of small cells.

Conclusion

This first-hand understanding has given researchers an advantage in network planning and in optimizing the customer experience. “We went into the street, did the testing, and we know how small cells work – that’s a confidence our customer’s should have,” said Puthenpura.

The state-of-the-art HARP tool was built based on this research, under the leadership of Raj Savoor, Assistant Vice President. The two teams worked very closely in creating HARP, paying special attention to reduce the computational complexities. HARP uses a programming model that places small cells in the most optimal position to maximize coverage and capacity for the network as a whole. Key data that the tool takes into account includes high-resolution geo-located demand, macro network RF and performance metrics, high-resolution clutter, 3D structure polygons, as well as building location level information. Transport availability and cost are evaluated by the algorithm further along in the process, during candidate selection. This automated capability enables RF planners to focus their attention on a smaller set of potential locations, allowing for more detailed design considerations.

The HARP tool has been highly instrumental in the planning phases of small cells and will continue to do so as we work to strategically enhance our network.

201311_Small_Cells_Fig_05_heatmap201311_Small_Cells_Fig_05_heatmap01

Figure: Heat Map

About the Author

Sarat Puthenpura is Executive Director of  Service Quality Management  Research, responsible innovating new capabilities and tools to manage end-to-end service quality.  He joined AT&T Bell Labs in 1986 and has worked in a number of areas in telecommunication, predominantly in network planning and optimization spanning both wireless and wireline networks; in particular applications of mathematical programming techniques to solve complex problems in these areas.  In addition to R&D activities, he has hands on experience in planning, implementing, and operating AT&T’s joint venture GSM network in India (currently Idea Cellular) from 1995 to 1998. Most recently, he has been leading the creation of an integrated network capacity planning and forecasting platform called NOVA (Network Optimization, Visualization and Analysis), which is extensively used for planning AT&T’s UMTS and LTE Radio Access Networks.

Sarat received B.Tech in Electrical Engineering from IIT Madras and Ph.D. in Electrical Engineering from McMaster University, Canada. In addition to several research papers, he has been the author of a graduate level textbook in mathematical programming ("Linear Optimization and Extensions - Theory and Algorithms", Prentice-Hall, 1991).  He has taught in the Department of Electrical Engineering, Industrial Engineering, and well as in the Graduate School of Management, Rutgers University. He is a Senior Member of IEEE and serves on the editorial board of the IEEE/ACM Transactions on Networking.

Current research interests:

  1. End to end service quality management  of VoLTE and video over wireless.
  2. Service quality management and tooling for cloud services.
  3. Service quality management under virtualized network infrastructure.
  4. Application of Big Data in service quality management.