The best monitoring hardware demands the best software — and this is it. Everything you need to manage the spectrum in real time, hunt interference, schedule missions, derive actionable intelligence and react to threats. New features and capabilities are added with every release.
Resolve complex interference in real time.
Many different signal types are potential sources of interference, indicators of illegal activity or unauthorised spectrum use.
Direction finding and individual geolocation techniques are each effective for only a limited range of target signal types, which varies according to many factors including signal bandwidth, modulation, power, background noise, pulse duration and receiver speed.
The most reliable and cost-effective method to ensure detection and successful localization of the maximum range of target signal types is to combine each technique in a single real-time overlay. The RFeye® supports multiple DF and geolocation techniques including Angle of Arrival (AOA), Time Difference of Arrival (TDOA) and Power on Arrival (POA).
Using machine learning techniques and an extensive training data set, CRFS has created a highly accurate signal classifier.
The RFeye® Signal Classification module is able to identify a comprehensive range of modulation types, both analog and digital, covering military and civilian standards. The module identifies the modulation technique of a signal of interest and returns it in string format. This gives you center frequency along with modulation type (“QPSK” or “QAM,” for example), the number of modulation levels, modulation index and baud rate. All of this helps to identify the type of transmission detected and can allow operators to diagnose causes of interference and identify possible threats.
The RFeye Signal Classification module has been developed using the latest Machine Learning techniques and mathematical tools to train neural networks and CRFS’s very extensive training data sets of signals. Machine learning techniques are ideal for optimizing signal classification performance and allow for very fast accurate signal classification while minimizing false alarms and permit easy updating of signal classifier databases with new signal types.
How does it work?
During the training process, the performance of the neural net can be monitored using a confusion matrix as shown below. The input data for training neural nets is a large database of signal types collected from real-world RF environments using CRFS-distributed RF sensors located all over the world. The more signal types, with real world distortions such as multipath, the more effectively the neural net can be trained to perform in actual field deployments.
Confusion matrix for signal recognition
The CRFS training database uses “tagged” signals with known classifications to train the neural net. The training process adjusts individual neurons in the net so that they can identify the known signals correctly, i.e., the classifier’s calculation and the “tag” match.
To test the neural network, a separate set of known signals not used for the training process is input to the classifier. The confusion matrix shows the known input signal type against the classifier’s calculation for every test and increments the appropriate element of the matrix with the answer against the test input. In other words, a 100% correct classifier would produce a confusion matrix with non-zero elements only along the leading diagonal. Generally, test scores of >95% accuracy can be achieved. These test scores improve with more varied and deeper training data sets. CRFS is actively building massive tagged-signal training databases for this purpose.
Typical signal classifier neural network
The input layer of a neural net has a series of weightings from features extracted from signal IQ data, e.g., 20 feature weightings and 20 input neurons. The output layer has a neuron for each signal type to be classified, e.g., 30 signal types and 30 output neurons. The neural net selects the output signal type to reflect the most probable signal given the input features.
Record and playback
The software allows data to be recorded and played back via a simple interface. Recording can be manual using the “record” button, and all data processes that are running in the software at the time will be recorded. Alternatively, recording can be set to trigger at a mask breakage, in which case all data relating to the mask breakage will be recorded using predetermined parameters.
The Propagation Analysis module is a software plug-in that can be used to accurately model wide-area spectrum monitoring capabilities at real-world locations.
The Propagation Analysis module offers immensely powerful RF simulation, planning and analysis tools for both natural topographical and man-made structural features.
The terrain data is first analysed for initial placement of receiver stations. This analysis allows the user to investigate the impingement on the first Fresnel zone between the receiver and potential target location. The analysis tools then use the built-in propagation models to investigate receiver and geolocation coverage. Most geolocation techniques require simultaneous data from one or more receivers, so a coverage analysis looks at how many receiver coverage areas can be seen at the same time. The receiver stations can then be best positioned to optimize the geolocation techniques being used.
The propagation of RF signals can be modeled to varying degrees of complexity. In the most basic form, the Free Space model can be used; this simply relies on the Friis Transmission Equation, allowing simulation of the decrease in received power for increasing frequency and distance from the transmitter. Complexity and real-world correspondence is cumulatively built up with Earth Curvature, Line of Sight and Fresnel models, allowing the inclusion of horizon effects, obstacle shadowing, and obstacle diffraction and reflection effects respectively. This propagation modeling ensures the accuracy of geolocation and monitoring simulations as well as allowing propagation-analysis coverage maps to be generated.
Receiver Nodes can be set to automatically upload spectrum measurements to a database. This allows remote viewing of Node data over IP as well as querying, filtering and graphing of the measurements of interest. Whether you’re organizing Nodes into user-defined groups or searching occupancy graphs by time and date, RFeye® software makes large networks, and the resulting data, easy to manage.
Other database functions for big spectrum data include management and filtering of historical spectrum sweeps for overlay onto spectrum maps and management of licenses with the licence database.