Memristor Applications

What Memristive applications are on the horizon, and how close are they to reality? We look at a survey of memristor applications and technology, starting from what the first devices will look like, and where they might go. This reference page will be updated as advances in each of the areas are made.

– Non-volatile memory applications: Memristors can retain memory states, and data, in power-off modes. Non-volatile random access memory, or NVRAM, is pretty much the first to-market memristor application we’ll be seeing. There are already 3nm Memristors in fabrication now. Crossbar latch memory (see below) developed by Hewlett Packard is reportedly currently about one-tenth the speed of DRAM. The fab prototypes resistance is read with alternating current, so that the stored value remains unaffected. Rosy colored industry analysts state there is industry concurrence that these flash memory or solid state drives (ssd) competitors could start showing up in the consumer market within 2 years.

– Low-power and remote sensing applications: coupled with memcapacitors and meminductors, the complementary circuits to the memristor which allow for the storage of charge, memristors can possibly allow for nano-scale low power memory and distributed state storage, as a further extension of NVRAM capabilities. These are currently all hypothetical in terms of time to market.

– Crossbar Latches as Transistor Replacements or Augmentors: The hungry power consumption of transistors has been a barrier to both miniaturization and microprocessor controller development. Solid-state memristors can be combined into devices called crossbar latches, which could replace transistors in future computers, taking up a much smaller area. There are difficulties in this area though, although the benefits these could bring are focusing a lot of money in their development. So perhaps the “where theres a will, or a dollar, theres a way” adage will get these developed. Unless a competition war amongst industry giants becomes one of those patent showdowns, where companies buy out technological advances to bury them. Remember 3G? Well, someone bought out 4G back in 2004, before 3G even came to market, and has been sitting on it ever since. And have profited greatly.

– Analog computation and circuit Applications: There was a track of electrical/mathematic engineering which was largely abandoned to stasis in the 1960s, as digital mathematics and computers rose to dominance. Analog computations embodied a whole area of research which, unfortunately, were not as scalable, reproducible, or dependable (or politically expedient in some cases) as digital solutions. However, there still exist some very important areas of engineering and modeling problems which require extremely complex and difficult workarounds to synthesize digitally: in part, because they map economically onto analog models. (see neuromorphic and learning circuits below.) The early work of Norbert Wiener has already started to be revisited, after the analog/digital split between him and John vonNeumann. Analog was great, but required management for scalability beyond what even the extremely complex initial digital vaccum tube computers could provide. Memristor applications will now allow us to revisit a lot of the analog science that was abandoned in the mid 1960’s.

– Circuits which mimic Neuromorphic and biological systems (Learning Circuits) – This is a very large area of research, in part because a large part of the analog science detailed above has to do with advances in cognitive psychology, artificial intelligence modeling, machine learning and recent neurology advances. The ability to map peoples brain activities under MRI, CAT, and EEG scans is leading to a treasure trove of information about how our brains work. But modeling a brain using ratiocinated mathematics is like using linear algebra to model calculus. Simple electronic circuits based on an LC network and memristors have been built, and used recently to model experiments on adaptive behavior of unicellular organisms. The experiments show that the electronic circuit, subjected to a train of periodic pulses, learns and anticipates the next pulse to come, similar to the behavior of the slime mold Physarum polycephalum periodic timing as it is subjected to periodic changes of environment. The recent memristor cat brain is also getting a lot of mention. These types of learning circuits find applications anywhere from pattern recognition to Neural Networks. No more neural pattern algorythm training on stock market data for the pop-sci investor: now, you can grow your own neural network! Just add two drops of memristor. Not anywhere close to reality, FYI, even in the 30 years range, but very realistic in terms of helping advance the science itself, if not the consumer market for intelligent brains-in-a-jar.

– Programmable Logic and Signal Processing, and a variety of Control System memristor patents are out there, waiting for the microchips to fall where they may. The memristive applications in these areas will remain relatively the same, because it will only be a change in the underlying physical architecture, allowing their capabilities to expand, however, to the point where their applications will most likely be unrecognizable as related.