Wireless Wanderlust Competition

In this year’s Data to Information (D2I) competition, we explore a semi-supervised indoor positioning challenge. Consider an indoor environment with multiple remote antennas, each having several antenna elements. These antennas operate over a set of transmission frequencies, capturing signals from users who carry communication devices. We refer to the relevant channel characteristics as Channel State Information (CSI). The CSI observations are transmitted from the remote antennas to a fusion center over a grants-free random access channel. Grant-free random access refers to a communication model in which devices transmit data directly, without requesting permission or waiting for explicit scheduling from a central controller. By analyzing the received signals at each remote antenna, a fusion center infers the user’s position.

Competition Dates: May 7th - June 13th, 2025

Registration deadline in...

Registration and Slack Links

Registration Link

--> Click here to register for the ISIT 2025 Wireless Wanderlust Competition <--

Slack Link

For general competition information and discussions, join our Slack channel here. If you are looking for team members, join the channel here.

ISIT 2025 Date to Information (D2I) Competition:

Wireless Wanderlust

Competition Dates: May 7th - June 13th, 2025

Registration Deadline: May 7th, 2025

The competition will use the DICHASUS Massive MIMO CSI Dataset from the Telecommunication Institute of the University of Stuttgart. Owing to the German origin of the dataset, we have termed the competition Wireless Wanderlust.

Objective: The competition tasks are threefold:

Vanilla Localization:
- Objective: Train a model that takes CSI estimates from four remote antenna arrays, each consisting of 8 elements, over 16 frequency bands, and produce location estimates for the user’s device.
- Evaluation: Localization Accuracy is evaluated based on the MAE and R90 metrics.
Trajectory-Aware Localization:
- Objective: Leverage the estimates of the user’s location in the previous timesteps, along with the CSI estimates, to increase localization accuracy.
- Evaluation: Localization Accuracy is evaluated based on the MAE and R90 metrics.
Feature Selection and Grant-free Random Access:
- Objective: Given a grant-free random access communication medium, design an opportunistic communication system and feature selection mechanism for user localization.
- Evaluation: Localization Accuracy is evaluated based on the MAE and R90 metrics.

Eligibility and Participation: The team can consist of up to three members, of which there should be at least two student members. At least one of the student members of the competing team should be registered as a member of the IEEE Information Theory Society at the time of registering for the competition.

Registration and Link to Training Data: Teams can register for the competition through the following Link during the registration window. The training data and corresponding loader programs will be made available through the Kaggle portal.

Team Formation: Each participant may register as a member of a team, or request to be teamed up with other participants by the competition organizers.

Submission of Results: The submission page can be accessed via the Kaggle portal. Additional information on how to access the portal will be provided through the ISIT website and ITSoc social media platforms prior to competition start date.

Tentative Prizes: