Multistatic Noncoherent Linear Complexity Miller Sequence Detection For Gen2 RFIDIoT
Abstract : Passive Gen2 radio frequency identification (RFID) tags work thanks to the utilization of line codes that balance their operation between two opposite states: absorbing RF (for energy harvesting) and reflecting RF (for backscattering/communications). Given the current RF harvesting technology, batteryless tags need to be located very close to an active illuminator in order to harvest sufficient energy and operate. To tackle this limitation, prior art has tried to bring the illuminator closer to the tags by designing proprietary illuminating architectures. Our solution comes in two parts. First, we offer a novel, Gen2-compliant, near-optimal, noncoherent sequence detection algorithm with linear complexity (in the sequence length) for Miller line codes. We leverage the robustness of this algorithm to overcome issues inherent in multistatic setups, such as carrier frequency offset. Second, we propose a modular multistatic architecture that makes use of low-cost commodity software defined radios (SDR) and omnipresent Ethernet infrastructure. Simulations and experimental results in a monostatic, bistatic, or multistatic SDR testbed with commercial RFIDs, corroborate the low-cost, real-time and near-optimal flavor of our solution.
? The co-operation of sensors and sharing of information in existing installations is a major focus which generally amounts to an spatially incoherent multistatic system.
? Such fluctuations will also exist for the transmitter-receiver pairs which make up a multistatic system, where the target is viewed from an array of different aspect angles.
? It is important to clarify in which situations a strong correlation between amplitude and phase values between receive sites are likely to actually exist.
? This correlation may exist at Doppler offsets and leads to a potential reduction in the effective level of phase noise where the delay time is short enough.
? One of the most challenging problems that needs to be solved in order to implement Ethernet/network-based multistatic RFID readers is the CFO compensation.
? The rest of the samples are compensated using the solution of a least squares-based, best linear fit problem, utilizing the stored phases of the PLL output.
? It can be investigated as to whether any of the problematic terms in L5 are removed with assumptions about received signals.
? However some simple examples can be considered where noise and average SNR is equal for target returns across all transmitter-receiver pairs; in which case the problem is identical to that of pulse integration.
• One often proposed advantage of multistatic radar, in that typically down-range information will be much more accurate than cross-range obtained from estimations of antenna azimuth and beamwidth, due to practical limitations on antenna size.
• However an overview of the proposed advantages of using multistatic radar is taken here, and mention given to some of the technical challenges and disadvantages to operating such a system.
• This was done firstly for purpose-built targets which could easily be simulated, before more complex targets such as a vehicles and people were investigated.
• Multistatic radar can be used for the purposes of distinguishing between real targets and multipath reflections.
? It is crucial to note that due to residual CFO (even after the PLL stage), this channel estimate is not robust and it deteriorates the performance of the detector in terms of bit error rate (BER).
? The limited performance of RTL-SDR is most likely caused by the large buffer size offered by its software driver, in conjunction with the low transfer speeds of USB 2.0, resulting to delays and missed blocks of samples.
? The system was tested using both coherent and noncoherent sequence detection schemes and the performance, with respect to tag reading rate was similar.
? Even though the results were acceptable, the PLL carrier tracking GNURadio block was used instead.
|