FlockLab: Wireless Link Quality Estimation Data

ETH Zurich's FlockLab testbed provides wireless link quality data collected from 27 to 29 sensor nodes. The dataset contains packet reception events (success/failure) from repeated tests run every two hours over several months on TelosB and DPP-cc430 platforms. It was created to support research into wireless network estimation and includes the data collection firmware for reproducibility.

Use Cases

• Modeling packet delivery ratios based on radio frequency channel and transmit power settings.
• Analyzing temporal stability of wireless links in a controlled testbed environment.
• Developing and benchmarking link quality estimation (LQE) algorithms using real reception logs.
• Studying the performance differences between TelosB and DPP-cc430 wireless platforms.

Strengths

• Includes data from two distinct hardware platforms (TelosB and DPP-cc430), enabling comparative analysis.
• Covers a multi-month time range, allowing for the study of temporal variations in link quality.
• The associated data collection firmware (Baloo-based) is provided, supporting reproducibility and extension to other networks.

Limitations

• Conflicting reports on dataset size: one source mentions 3 weeks with ~250 tests/platform, while another mentions 3 months with ~12 tests/day/platform and 5 months with ~4 tests/day/platform.
• Specific column names and data schema are not provided, complicating initial data understanding.
• Metadata such as exact row count, file size, and last update date are unknown.

Provenance

Source

ETH Zurich, FlockLab testbed

Collection Method

Controlled experiments where nodes sent 100-packet strobes in dedicated time slots, with all other nodes logging reception events.

Time Range

Multiple months (specific range conflicted across sources)