Comb Decoding towards Collision Free WiFi
summary
CombDec treats RTS frames as vectors with a comb-like distribution and reformulates RTS collisions as a sparse recovery problem, using constrained RTS fields (frame control, addresses, duration, sequence/LAV, FEC) to prune the search space. Implementation on USRP builds a comb matrix from successfully decoded RTS samples, runs compressive-sensing style decoding (alphabetic construction + decimation) to recover colliding RTS (primarily two transmitters), and reports 33.6%–46.2% throughput gains and high collision-resolution probability in experiments.
AI-written summary · about
chapters (8)
- 0:00Start
- 0:56RTS/CTS based Channel Sensing
- 1:48Collision Resolution
- 5:36WiFi Packet Trace Analysis
- 8:47CombDec Architecture
- 10:08Single Link Evaluation
- 10:46Network Evaluation
- 11:45Summaries
related talks
19mNetScatter: Enabling Large-Scale Backscatter Networks
13mVerification and Redesign of OFDM Backscatter
19mOrthcatter: High-throughput In-band OFDM Backscatter with Over-the-Air Code Division
15mDecoding RSSI Compression in RFID: Dynamic RCS Modeling and Tag-Intrinsic Power Metrics..
28mVLBI with GNU Radio and White Rabbit
12mToward Nearly-Zero-Error Sketching via Compressive Sensing
22mInference Deployments and Comms Implication
18mRetwork: Exploring Reader Network with COTS RFID Systems