Bastille’s Differentiating Technology
To protect organizations against existing threats and new threats emerging via the radio spectrum there are 3 areas of technology across which Bastille has 17 patents approved and pending. These technologies permit Bastille to SENSE, IDENTIFY & LOCALIZE threats as follows:
Collaborative Bandit Sensing is Bastille's patent pending technology (USPTO APP # 20160127404) to quickly and accurately scan the spectrum for emitters and threats. Collaborative Bandit Sensing utilizes a variant of the Multi-Armed Bandit (MAB) problem to allocate sensor time watching various parts of the spectrum. The classic MAB problem is one where an actor has to balance exploitation of a reward function with exploration of other possible activities that might lead to a higher payoff. The challenge is that exploration has opportunity costs, while premature exploitation can lead to a low payoff rate. In Bastille’s context, our sensors are intelligently making distributed decisions about when they should continue to observe a known signal (exploitation) versus scanning another part of the spectrum to find unknown signals (exploration). Bastille has solved this problem by combining a stochastic model of the signal environment, an array of intelligent distributed search algorithms, and collaborative optimization based on gossiping algorithms.
Bayesian Device Fingerprinting is a suite of patent pending technology (USPTO APP # 20150348341) that Bastille uses to detect and identify devices in an Enterprise airspace. Bayesian Device Fingerprinting relies on Probabilistic Graph Models of possible device characteristics to track and estimate device meta-information. The probability states are updated as new observations are made both in radio frequency space and through other input systems like access control, MDM, video surveillance, or SSO. With Bayesian Device Fingerprinting, Bastille Enterprise can resolve emitter, device, and people-device entities to produce never-before-seen situational awareness of your RF and physical space.
Distributed Tomographic Localization is Bastille's patent pending technology (USPTO APP # 20160127931) to provide actionable position information of all emitters in your corporate airspace. Like Computer Tomography (CT scans) uses path loss in the medical world, Bastille uses passive Radio Tomography in the corporate airspace to account for the location of walls. This allows Bastille to locate radio emitters much more accurately than other technologies to 1 metre of accuracy.
Medical CT scans send X-Rays (electromagnetic or EM signal) beams at various angles through a cross section of your body. The attenuation of each beam is collected and processed to create a cross-sectional image of your torso through tomographic imaging algorithms. The image produced is an image of the X-Ray Spatial (path) Loss Field (SLF) of your organs. Because your organs have varying compositions, they attenuate EM energy differently, leading to a useful image or your organs.
Similarly Bastille uses the passive RF signals to "image" an office space and estimate the space's SLF. The resulting SLF tells us where attenuating objects, like walls, are in the environment. When Radio Frequency (RF) signal are used, as opposed to X-rays, the process is called Radio Tomographic Imaging (RTI).
Bastille’s approach to localization uses two unique innovations to achieve industry-leading location accuracy. The first is Distributed Tomography, which allows Bastille to estimate the positions of the walls and other objects in the environment which we then incorporate into our localization model. The second is Bayesian filtering coupled with Ensemble Machine Learning algorithms to do precise perimeter detection. Bastille Enterprise customers can use Distributed Tomographic Localization to geo-fence emitters and set localization based alerts for sensitive areas.