Specialty Processing
TrellisWare continues to develop unique processing approaches to boost the performance of both standalone applications as well as enhancements to our own products. These emerging concepts take advantage of our knowledge of signal processing and efficient communication design – from direction finding to blind modulation classification and signal acquisition to blind separation of multiple signals with unknown timing relationships. By “blind,” we mean without a prior knowledge of things like signal BW, baud rate, synchronization, SNR, or embedded training sequences. While our technology mantra is to use all the knowledge available, some amazing things are possible using TrellisWare Likelihood Processing (TLP) when very little is known.
Direction Finding
Unlike conventional interferometry or matrix-based DF approaches, TrellisWare’s approach to Direction Finding (DF) and location finding uses the knowledge of the structure of the signal of interest and the expected time dispersal of the channel in our DF processing. From this approach TrellisWare has derived patented techniques for determining more Angles of Arrival (AOA) than antennas used, something that is mathematically impossible with traditional approaches. These techniques also allow TrellisWare solutions to perform DF processing in the midst of bad reflections and/or interference and at very low SNRs (or very short measurements or records).Thus, we can work in environments where typical approaches fail and derive information otherwise unattainable.
The TrellisWare approach can be thought of as deriving a “spatial matched filter” (SMF) based on input from multiple antennas that contains inherent spatial information. From this SMF, various parameters of interest can be derived, including relative AOA and TOA of the signal and its reflections as well as measures of SNR and confidence per reflection. The SMF also allows operation in much lower SNR than usual, so TrellisWare processing can be used on very short sets of data, in many cases allowing processing on single bursts for TDMA-type signals.
Applications for this specialty processing technique include: commercial cellular E911 applications when GPS is not available or in conditions where triangulation from the tower is not sufficient; ground DF in high multipath urban areas; and network topology awareness for TrellisWare’s own mobile mesh solutions.
Cognitive Radio Techniques
Radios that can autonomously and efficiently characterize the spectral environment blindly are of increasing importance for “cognitive” commercial tasks such as spectrum sharing in short-term leased bands. When done in a non-cooperative manner, many parameters and features must be identified quickly, efficiently and accurately. TrellisWare has developed Modulation Classification (MC) techniques that render efficient decisions on the exact transmit modulation type out of a given menu of possibilities. Techniques such as TLP provide outstanding results for blind MC against higher throughput modulations operating in time-varying, multipath-affected, congested channels.
TrellisWare also is working on a number of methods for acquiring signals of known structure, but unknown data, timing and exact frequency/Doppler in highly demanding environments. With so many unknowns, algorithms such as Per Survivor Processing (PSP) are essential to accurately acquire the signal. In many cases, after such blind acquisition the signal can be tracked and demodulated. In fact, if the signal is not encrypted and uses error correction coding, adaptively iterating between the decoded data and the blind estimation parameters (AID) can often provide even greater performance gains.
Co-Channel Interference (CCI) Mitigation & Single Antenna Interference Cancellation (SAIC)
Coping with co-channel interference is vital to increasing network capacity in congested spectral channels. High levels of frequency reuse increases co-channel interference problems, and make co-channel interference the limiting factor in increasing network capacity.
TrellisWare has developed an innovative Single Antenna Interference Cancellation (SAIC) algorithmic approach that resides in the digital baseband processing section of the receiver that allows handsets to operate under very strong interference conditions. TrellisWare’s SAIC allows significant increases in the number of subscribers per a given sector and hence network capacity.
Based on TrellisWare’s proven TLP techniques, TrellisWare’s SAIC equalizer is a drop-in replacement for the existing channel equalizers assuming the responsibilities of the equalizer it replaces and adds SAIC capabilities. The TrellisWare SAIC Equalizer can either be implemented fully in hardware, fully in software, or partitioned into a hybrid hardware/software implementation.
TrellisWare has also developed multiantenna versions of this CCI mitigation approach. While multiple antennas and receive chains imply increased complexity, even greater performance is possible by using the spatial information jointly with the processing used for SAIC. Now multiple signals and interference can be sorted not only based on timing and signal structure but also on angle of arrival.
TrellisWare's unique CCI mitigation AID techniques can separate multiple, simultaneous signals of exactly the same type. This graph illustrates that with sufficient iteration and SNR, multiple overlapping signals are demodulated with the same performance as a single signal.
The portable AOA test bench shown here is used for a variety of multi-antenna evaluations and field tests including terrestrial AOA testing.
TrellisWare's blind modulation recognition algorithms can distinguish a complex mix of signal types, as the decision probabilities for the frequency non-selective random phase channel show here.
Related Materials
- PSP Based Modulation Classifier
(246KB PDF)
