WCSNG @ UC San Diego

Code

  • WiROS: WiFi sensing toolbox for robotics

    WiROS makes three concrete contributions, in order to deliver an accurate, versatile and tractable WiFi-sensor framework for ROS. First, it provides a C++ framework to integrate WiFi-sensor specific API’s into ROS. We accomplish this for Nexmon API, however, newer platforms can be similarly integrated with little additional effort. Second, we provide a wireless calibration algorithm and toolkit to measure and calibrate for hardware non-idealities for WiFi sensors. The lastly, we open-source state-of-art algorithms to extract physical parameters like angles of arrival or departure of the WiFi signals.

  • mmFlexible: Flexible Directional Frequency Multiplexing for Multi-user mmWave Networks

    Our dataset which consists of multiple indoor and outdoor experiments for up to 30 m gNB-UE link. In each experiment, we fixed the location of the gNB and move the UE with an increment of roughly one degrees. The table above specifies the direction of user movement with respect to gNB-UE link, distance resolution, and the number of user locations for which we conduct channel measurements. Outdoor 30 m data also contains blockage between 3.9 m to 4.8 m. At each location, we scan the transmission beam and collect data for each beam. By doing so, we can get the full OFDM channels for different locations along the moving trajectory with all the beam angles. Moreover, we use 240 kHz subcarrier spacing, which is consistent with the 5G NR numerology at FR2, so the data we collect will be a true reflection of what a 5G UE will see.

  • BSMA: scalable LoRa networks using full duplex gateways

    The MATLAB simulator allows simulation of LoRa networks with various MAC protocols. The physical layer is abstracted out and the simulator can be used to evaluate the performance of MAC protocols in different network topologies.

  • Two beams are better than one: Towards Reliable and High Throughput mmWave Links

    This repository contains the artifact for submission #441, ACM SIGCOMM 2021. The artifact is composed of simulations and algorithms implemented on real-life mmWave channel estimates.

  • Pointillism: Accurate 3D Bounding Box Estimation with Multi-Radars

    This is the official code release for RP-net. It is the deep-learning system of Pointillism which estimates 3D bounding boxes from Cross-Potential point clouds generated by Pointillism.

  • Deep Learning based Wireless Localization for Indoor Navigation

    While being the first in in Deep Learning based Indoor Navigation with WiFi data, we want to build WiFi CSI dataset on par with ImageNet to assist further research in WiFi based indoor localization and their applications.

  • SIGNet: Semantic Instance Aided Unsupervised 3D Geometry Perception

    Unsupervised learning for visual perception of 3D geometry is of great interest to autonomous systems. This paper introduces SIGNet, a novel frameworkthat provides robust geometry perception without requiring geometrically informative labels. Specifically, SIGNet integrates semantic information to make unsupervised robust geometric predictions for objects in low lighting and noisy environments. SIGNet is shown to improve upon the state-of-the-art unsupervised learning for geometry perception by 30% (in squared relative error for depth prediction). In addition, SIGNet improves the dynamic object class performance by 39% in depth prediction and 29% in flow prediction.

  • SparSDR: Sparsity-proportional Backhaul and Compute for SDRs

    SparSDR’s goal is to make SDRs capture primary transmissions rather than entire channels. While a Full-capture SDR always backhauls data at a fixed rate, SparSDR takes advantage of frequency-time signal sparsity to scale the backhaul rate linearly with the actual occupancy of the channels observed. This allows SparSDR to backhaul more than 100 MHz of bandwidth over a backhaul where a Full-capture SDR could do less than 25 MHz.

  • SweepSense: Sensing 5 GHz in 5 Milliseconds with Low-cost Radios

    We propose a new receiver architecture for spectrum sensing radios where sampling is done along with quick sweeping of the center frequency. This is motivated by the intuition that a sweeping radio may miss lesser transmissions than one that sequentially tunes to different bands. We implement this using an open loop VCO fed with a sawtooth voltage waveform. The output of the VCO is used to drive a mixer and implement the sweeping radio. The architecture has been prototyped on a USRP N210 with a CBX daughterboard. Downconverting while sweeping introduces distortions in the signal, which we remove using an 'unsweeping' process and is discussed in the paper.

Datasets

  • mmFlexible: Flexible Directional Frequency Multiplexing for Multi-user mmWave Networks

    Our dataset which consists of multiple indoor and outdoor experiments for up to 30 m gNB-UE link. In each experiment, we fixed the location of the gNB and move the UE with an increment of roughly one degrees. The table above specifies the direction of user movement with respect to gNB-UE link, distance resolution, and the number of user locations for which we conduct channel measurements. Outdoor 30 m data also contains blockage between 3.9 m to 4.8 m. At each location, we scan the transmission beam and collect data for each beam. By doing so, we can get the full OFDM channels for different locations along the moving trajectory with all the beam angles. Moreover, we use 240 kHz subcarrier spacing, which is consistent with the 5G NR numerology at FR2, so the data we collect will be a true reflection of what a 5G UE will see.

  • Two beams are better than one: Towards Reliable and High Throughput mmWave Links

    This repository contains the artifact for submission #441, ACM SIGCOMM 2021. The artifact is composed of simulations and algorithms implemented on real-life mmWave channel estimates.

  • Pointillism: Accurate 3D Bounding Box Estimation with Multi-Radars

    This is the official code release for RP-net. It is the deep-learning system of Pointillism which estimates 3D bounding boxes from Cross-Potential point clouds generated by Pointillism.

  • mMobile: Building a mmWave testbed to evaluate and address mobility effects

    We release 28 GHz full channel (CSI) measurements for a mobile user. The CSI data is tagged with each user location and for each beam index. The CSI consists of 256 subcarriers with a sub-carrier spacing of 240 kHz requisite by 5G NR standards. There are four datasets for various indoor and outdoor environments.

  • Deep Learning based Wireless Localization for Indoor Navigation

    While being the first in in Deep Learning based Indoor Navigation with WiFi data, we want to build WiFi CSI dataset on par with ImageNet to assist further research in WiFi based indoor localization and their applications.

Search