Research
Research
Embedded Files
Spatial Oversampling for Robust Short-Range LoS MIMO
Spatial Oversampling for Robust Short-Range LoS MIMO
LoS MIMO links in the tera-Hertz band can be used to provide high speed point to point for emerging picocellular networks. In such settings, geometric misalignment between transmit and receive arrays is inevitable Furthermore, link ranges may vary significantly in such systems (e.g., from 50 to 150 meters in an urban backhaul). Geometric misalignment leads to a delay spread across the receive aperture causing ISI and CSI. Further, operating at ranges smaller than the nominal leads to mode collapse for fixed apertures. We propose an architecture for a joint adaptive space-time equalizer with spatial oversampling by introducing additional receive antennas while maintaining symbol-rate sampling to make the system robust to both geometric misalignment and link range variations. We consider linear space-time equalization and discuss a novel adaptive windowing scheme to control the complexity of the equalizer.
Read the full conference paper here!
Read our journal paper (submitted to IEEE TWC) here!
Fig. 1 : Example deployment scenario where misalignment of panels and varying link ranges of operation is an inevitable consequence
Fig. 2 : The proposed reflect-array aided LoS MIMO system for long-ranges
Reflect-arrays for Long-Range LoS MIMO Links
Reflect-arrays for Long-Range LoS MIMO Links
LoS MIMO is by now well accepted as a means of taking advantage of the increased spatial degrees of freedom available as we scale up the carrier frequency. However, for fixed transceiver form factors, the available degrees of freedom decay quickly with link range. We explore an innovative approach for creating spatial degrees of freedom for long-range links by utilizing large virtual apertures created by placing multiple reflect-arrays near the transceivers. Placing the reflect-arrays close to the transceivers necessitates beam-focusing for the transmit subarray and corresponding reflect-array link. Thus, the effective phase compensation required by each reflect-array on the transmitter side is two fold, accounting for the quadratic phase compensation from beam-focusing and linear phase compensation required to beamform towards the reflect-arrays on the receiver side. The use of reflect-arrays in this setting provides beamforming gain in addition to creating spatial degrees of freedom while retaining compact transceiver form-factors, enabling power efficient transfer of 5.2 Gbps over 1,500 m in the 28 GHz band.
View the full conference paper here!
mmSnap: Distributed Radar Sensing with Self-Calibration and One-Shot Fusion
mmSnap: Distributed Radar Sensing with Self-Calibration and One-Shot Fusion
Reliable situational awareness in urban environments requires sensing frameworks that can detect and track multiple moving targets beyond the field of view of any single sensor. In this work, we present mmSnap, a collaborative mmWave radar network that combines self-calibration with Bayesian one-shot fusion to enable scalable, low-latency sensing across distributed nodes. The system first self-calibrates opportunistically, aligning radar nodes relative to a reference using overlapping target tracks. With calibration in place, a Bayesian fusion algorithm integrates distributed measurements into instantaneous position and velocity estimates that match the accuracy of smoothed multi-frame tracking. We validate mmSnap through outdoor experiments using off-the-shelf 77 GHz MIMO radar hardware, demonstrating its potential for real-time, infrastructure-supported perception and monitoring in complex environments.
You can read the full paper here!
Fig. 3 : A collaborative networked RF sensing infrastructure.
Page updated
Google Sites
Report abuse