About Our Technology

This section will summarize a few techniques that ACRi has invented, refined, and/or demonstrated. We will update this section periodically with the latest algorithm breakthroughs or over-the-air (OTA) result analysis.

Non-Eigen Decomposition Beamforming

Many systems are unable to utilize adaptive blind beamforming due to their high computational needs, which is why ACRi members developed a new paradigm for beamforming that does not rely on the eigenvalues and eigenvectors used by conventional techniques. By removing all of the standard assumptions made by researchers in this field, Dr. Okamoto and Dr. Chen were able to invent a new class of smart antenna algorithms that significantly reduce the computational requirements while still achieving near-optimum adaptive blind beamforming performance.

Chen and Okamoto originally introduced the NED algorithm at the 2004 IEEE Vehicular Technology Conference. The algorithm has a computational complexity of just O(4M-4); in layman’s terms, that means that NED’s computational requirements are linear with the number of antennas (M), while standard algorithms require computational complexity proportional to the square or the cube (or much more!) of the number of antennas. Despite its ultra-low computational requirements, NED was shown to achieve comparable performance with standard algorithms that can require orders of magnitude more computations. This is even more impressive when taking into account that NED is a blind algorithm (no information on the user or interference location or spatial information required) while some of the algorithms NED was being compared against were non-blind and required that the spatial information or user and interference location be known in advance.

ACRi has significantly advanced the NED algorithm under SBIR funding from the National Science Foundation (Grant IIP-0810790) and will present a paper on these advances at the 2008 IEEE MILCOM Conference on November 18, 2008.  The key advance is the significantly improved interference capabilities for the NED algorithm with no decrease in algorithm performance. NED was originally designed under the assumption that the user signal was significantly stronger than all interference signals, which was the same assumption that competing low complexity algorithms made (for example, the RED algorithm developed by Choi and others that has computational complexity of M² instead of M like NED). Under ACRi’s NSF funding, Dr. Okamoto was able to modify NED to enable interference to be comparable to the user power, with simulations showing near-optimum performance despite the interference being stronger than the user signal. This is up to a 100x improvement in NED’s interference rejection capability with no decrease in performance.

High-Power Interference Mitigation

The proliferation of commercial wireless communications systems results is an ever-increasing problem of strong co-channel interference preventing successful radio communications. Military systems also use communications and surveillance equipment which cause mutual interference. Indeed, strong interference can come from nearby commercial, amateur, and military transmitters. While current state-of-the-art solutions to mitigate these problems had some success in the past, the rapidly increasing number of interfering transmitting devices is overwhelming these limited approaches. Consequently, next-generation communication systems require advanced techniques such as smart antenna beamforming to enable communications in the presence of strong in-band interferers. One example of this problem is how military radios have difficulty communicating in the presence of counter-IED jamming systems (which disrupts all communications including those by friendly radios) and on Navy ships (which has a difficult electromagnetic interference environment that makes high-speed reliable communications difficult).

Dr. Okamoto’s book shows over-the-air results from testbeds at three different frequencies of operation (800 MHz, 900 MHz, 1.9 GHz), as well as theoretical analysis and computer simulation results that are in agreement with the OTA results. OTA results showed that interference signals could be nulled 60 dB in stationary scenarios and numerous smart antenna algorithms were compared via their OTA performance in stationary and mobile scenarios. This implementation experience and systems design is directly applicable to tactical scenarios such as the counter-IED and shipboard environments.

Dr. Okamoto was the Principal Investigator for an ONR Phase II program on “MIMO Techniques for LPI/LPD/AJ Communications in Highly Mobile Networks.”  For this ONR-funded program, Dr. Okamoto designed a new adaptive blind beamforming algorithm that required significantly less computational complexity than standard algorithms while still enabling reliable communications even when an interference source was present that was 20 dB or more above the user signal. The OTA results closely agreed with the simulation results due to the realistic modeling used. The beamforming solution was demonstrated OTA to representatives from ONR and SPAWAR in 2007 and published at the 2007 IEEE Antennas and Propagation Society Conference.

ACRi members have advanced Dr. Okamoto's past work to provide even greater interference mitigation capability for high-power interference signals, with simulation results exceeding all of that previous work.  Further advances in the solutions and an OTA demonstration of the work depends on funding availability, with proposals pending.

Open-Source Hardware Prototype Testbed

Evaluating solutions over-the-air (OTA) in realistic scenarios is an essential ingredient for developing superior solutions. ACRi has developed our low-cost open-source testbed to quickly evaluate algorithms in real-world scenarios. This Software Defined Radio (SDR) hardware prototype is a COTS (Commercial-off-the-shelf) platform that includes an open-source RF front end, data acquisition car and PC interface, with glue logic via its FPGA. This modular system reconfigures with transceiver daughter cards for signals ranging from audible to about 3 GHz.

The key feature of this ACRi’s SDR prototype is reconfigurability. This cost-effective platform supports multiple communications waveforms and allows ACRi to quickly test solutions OTA in realistic situations with minimal effort and expense. The open-source nature of our platform means that its entire design is freely available under the GNU General Public License, allowing us to quickly modify even its most fundamental aspects to meet our evolving needs without having to replace hardware or software that could cost significant time and money. ACRi has already implemented several waveforms in the prototype and the radios already can transmit and receive signals.

Dynamic Slot Allocation

Dynamic slot allocation is a breakthrough technology to improve the capacity of wireless communication systems. Conventional TDMA (Time Division Multiple Access) systems (GSM, WiFi, etc.) only allow one user per frequency band in a time slot. SDMA (Space Division Multiple Access) enables multiple co-channel simultaneous users, where 8 simultaneous users sharing a frequency band can provide 8x capacity improvement. CDMA (Code Division Multiple Access) system capacity is interference limited and it has already been shown that smart antenna systems can significantly improve the system’s SIR (Signal-to-Interference Ratio), which increases the capacity for CDMA (one published example showed a 5x capacity improvement).

Dynamic slot allocation picks users to share time slots in each frequency band that result in minimal interference with the other users sharing the frequency band. This results in a significant improvement in capacity without a degradation in communications quality. This is far superior to current techniques that just picks users randomly for each time slot. Other dynamic slot allocation techniques have been published, but they require high computational complexity due to the costly computations of SINR and matrix inverses.

ACRi members have been working on the next generation of their Modified First Fit (MFF) technology for the past 6 years. MFF avoids the need to compute SINR because it uses ACRi’s Effective Relative Angle (ERA) measure to determine which time slot a user with its spatial profile should be added to in order to minimize system interference. ACRi has shown that a smart antenna system using ERA can provide between M/2 to 2M/3 times the throughput of a conventional system, where M is the number of antennas at the receiver.

MFF is an ultralow computational complexity technique that avoids the need to compute the costly matrix inverse. When compared to the published First Fit dynamic slot allocation technique (using MSINR), MFF requires only 388 flops while First Fit requires 11,076 flops!