Spectrum Challenge

The call for spectrum challenge submissions is now closed. You can visit our Spectrum Challenge Program page http://dyspan2017.ieee-dyspan.org/program/spectrum-challenge-program to see a list of accepted participants.

 

Introduction

The increase of heterogeneous networks requires novel techniques to learn and adapt to dynamic situations. The previous IEEE DySPAN focused on creating a highly spectral efficient radio without interfering with existing wireless technologies in the same frequency band. Inspired by the success of the previous challenge, we are moving on with an evolution of that challenge. This year, there is more focus on the learning and dynamic situational awareness. Can you make a really smart radio that knows perfectly in which situation it is, and adapt?

The IEEE DySPAN 2017 Spectrum Sharing Challenge is a competition designed to demonstrate a radio protocol that can achieve high spectral efficiency in a dynamic environment. The challenge will be setup as a competition between your radio protocol and a given opponent’s transmitter and receiver in an open-air conference environment in the 3190-3200 MHz band. Transmitter and receiver information from a known primary user will be logged into a database, allowing you to tune the sharing performance or your radio protocol. Participants can be experts on hardware, MIMO or beamforming solutions, novel waveforms, channel hopping MAC protocols, self-organizing networks or cognitive radio and networking in general. The only given is directions for accessing the database to obtain performance measurements and legacy receiver performance feedback. Different from last year, we will also have a PU receiver with multiple locations, so that non-adaptive beamforming is not going to be very useful any more.

Eligibility and Application

The challenge is open to any academic institution, business or individual entity. An applicant may be an individual, or representing an institution. Only one application per team should be submitted, and only one application per legal entity will be accepted for the challenge. The person identified as team leader should have a master degree. Individuals may be members of only one team.

Participation and awards

Interested contestants should submit a maximum 2-page paper to the DySPAN 2017 spectrum challenge track detailing the proposed method and some early results about their implementation by December 19, 2016 (Visit https://edas.info/ to submit a paper).

The best will be selected for the final competition.

Selected teams will also be invited to submit a final challenge paper to be included in the IEEE DySPAN 2017 proceedings and included in the IEEE Xplore database.

Two winners will be selected: a winner obtaining the best performance on agile spectrum use, and a winner with the best situational awareness.

Application Submission Deadline: December 19, 2016.
Notification of acceptance to competing teams: December 20, 2016.
Challenge paper camera ready (will be included in IEEE DySPAN demo conference proceedings): February 2017
Final event and selection of 2 winners: March 2017 (IEEE DySPAN)

Figure 1: Timeframe for the challenge before and @ DySPAN
Figure 1: Timeframe for the challenge before and @ DySPAN

 

Setup

The challenge is open to all kinds of innovations that would improve the PU receiver focused spectrum sharing, including novel waveforms, MIMO or beamforming solutions, novel hardware (fixed frequency and bandwidth), spectrum sensing, or higher layer learning schemes that improve how the SU adapts to the varying PU.  To achieve this, following items are predefined as also represented in Fig. 2:

  • SU antenna locations (2x2);
  • PU system with multiple Rx antenna locations;
  • Database delivering PU receiver feedback on throughput and packet loss, as well as delivering packets to the SU and monitoring the SU performance (documented API).

All teams will receive all necessary software to test their solution on a USRP X300. During the DySPAN challenge, the PU system will also be running on a software defined radio platform. A prototype PU software will be provided for SDR platforms for testing. The actual hopping/occupancy patterns might vary during the challenge. During the challenge, the PU will switch between 4 receiver antenna locations to simulate mobility. In addition to providing PU receiver feedback, the database will deliver packets to the SU and take care of the performance monitoring during the challenge. Fixed length layer-2 packets will be provided to the SU by the database, and the construction of layer-1 packets is up to the contestant.

Source code of the PU reference design, database access and API can be found here: http://claws.be/spectrum-challenge/.

Figure 2

The challenge will consist of two phases. In phase one the situational awareness of the teams is tested. The teams will need to detect the correct situation, i.e. set of PU parameters:

  • Bandwidth = {2.5, 5, 10} MHz
  • Carrier frequency + hopping pattern
  • Packet length = {100, 1000} bytes
  • Inter arrival time between packets

This phase will run for 10 minutes and consists of 5 situations, which are randomly selected from a superset, each lasting 2 minutes. The whole set of scenarios will be made public before the challenge. An example set of scenarios are given below. The teams need to detect as soon as possible the correct situation and transmit this to the database.


Scenario


Bandwidth

Carrier frequency

Packet length

Inter arrival time

1

5

Channel 2

100B

5ms (deterministic)

2

5

Hopping from channel 1 to 2

100B

5ms (deterministic)

3

10

Channel 3

1000B

2ms (deterministic)

4

10

Channel 2

1000B

Back-to-back

In phase two the teams will need to show how agile their SU is. During a second phase, the SU performance will be logged in the database, and winner will be selected based on those logs.

Figure 3

 

Winners

From the competing teams, two winners will be selected, one metric scores the teams in function of the throughput and one metric scores them in function of their spectrum situation awareness.  All out of band transmission will be heavily penalized (spectrum use will be monitored and filters will be provided).  Throughput is measured as total number of packets delivered to the database by the receiver during the challenge. The best team improves the product of its throughput and the satisfaction of the PU. As PU packets are dropped, PU satisfaction decreases rapidly. More information about the metrics can be found in the figure below.

Figure 4: Agile Spectrum Agility Use and Spectrum Awareness

 

Awards

Each winning team will be awarded a prize of $1,500USD.

 

Getting started

To get started, visit this webpage [http://dyspan2017.ieee-dyspan.org/spectrum-challenge/getting-started].