In one embodiment, the present invention provides a system for autonomous spectrum management in an electromagnetic environment including at least one monitoring sensor operable to autonomously monitor the electromagnetic environment and create measured data based on the electromagnetic environment, at least one data analysis engine for analyzing the measured data, a semantic engine including a programmable rules and policy editor, and a tip and cue server, wherein the at least one data analysis engine includes a detection engine and a learning engine, wherein the detection engine is operable to automatically detect at least one signal of interest, wherein the learning engine is operable to learn the electromagnetic environment, wherein the programmable rules and policy editor includes at least one rule and/or at least one policy, wherein the tip and cue server is operable to use analyzed data from the at least one data analysis engine to create actionable data, and wherein the tip and cue server and the at least one data analysis engine are operable to run autonomously without requiring user interaction and/or input.

In another embodiment, the present invention provides a system for autonomous spectrum management in a radio frequency (RF) environment including at least one monitoring sensor operable to autonomously monitor the RF environment and create measured data based on the RF environment, at least one data analysis engine for analyzing the measured data, a semantic engine including a programmable rules and policy editor, and a tip and cue server, wherein the at least one data analysis engine includes a detection engine, a geolocation engine, and a learning engine, wherein the detection engine is operable to automatically detect at least one signal of interest, wherein the learning engine is operable to learn the RF environment, wherein the programmable rules and policy editor includes at least one rule and/or at least one policy, wherein the semantic engine is operable to create a semantic map including target data, wherein the tip and cue server is operable to use analyzed data from the at least one data analysis engine to create actionable data, and wherein the tip and cue server and the at least one data analysis engine are operable to run autonomously without requiring user interaction and/or input.

In yet another embodiment, the present invention provides a method for autonomous spectrum management in an electromagnetic environment including providing a semantic engine including a programmable rules and policy editor, wherein the programmable rules and policy editor includes at least one rule and/or at least one policy, monitoring the electromagnetic environment using at least one monitoring sensor and creating measured data based on the electromagnetic environment, analyzing the measured data using at least one data analysis engine, thereby creating analyzed data, wherein the at least one data analysis engine includes a detection engine and a learning engine, learning the electromagnetic environment using the learning engine, automatically detecting at least one signal of interest using the detection engine, and creating actionable data using a tip and cue server based on the analyzed data from the at least one data analysis engine, wherein the tip and cue server and the at least one data analysis engine are operable to run autonomously without requiring user interaction and/or input.

These and other aspects of the present invention will become apparent to those skilled in the art after a reading of the following description of the preferred embodiment when considered with the drawings, as they support the claimed invention.

BRIEF DESCRIPTION OF THE DRAWINGS

The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fee.

FIG. 1 illustrates one embodiment of an RF awareness and analysis system.

FIG. 2 illustrates another embodiment of the RF awareness and analysis system.

FIG. 3 is a flow diagram of the system according to one embodiment.

FIG. 4 illustrates the acquisition component of the system.

FIG. 5 illustrates one embodiment of an analog front end of the system.

FIG. 6 illustrates one embodiment of a radio receiver front-end subsystem.

FIG. 7 continues the embodiment of the radio receiver front-end shown in FIG. 6.

FIG. 8 is an example of a time domain programmable channelizer.

FIG. 9 is an example of a frequency domain programmable channelizer.

FIG. 10 is another embodiment of a programmable channelizer.

FIG. 11 illustrates one embodiment of a blind detection engine.

FIG. 12 illustrates an example of an edge detection algorithm.

FIG. 13 illustrates an example of a blind classification engine.

FIG. 14 illustrates details on selection match based on cumulants for modulation selection.

FIG. 15 illustrates a flow diagram according to one embodiment of the present invention.

FIG. 16 illustrates control panel functions according to one embodiment.

FIG. 17 illustrates one embodiment of an RF analysis sub-architecture of the system.

FIG. 18 illustrates one embodiment of a detection engine of the system.

FIG. 19 illustrates a mask according to one embodiment of the present invention.

FIG. 20 illustrates a workflow of automatic signal detection according to one embodiment of the present invention.

FIG. 21 illustrates components of a Dynamic Spectrum Utilization and Sharing model according to one embodiment of the present invention.

FIG. 22 illustrates a Results model provided by the system according to one embodiment of the present invention.

FIG. 23 is a table listing problems that are operable to be solved using the present invention.

FIG. 24 illustrates a passive geolocation radio engine system view according to one embodiment of the present invention.

FIG. 25 illustrates one embodiment of an algorithm to select a geolocation method.

FIG. 26 is a diagram describing three pillars of a customer mission solution.

FIG. 27 is a block diagram of one example of a spectrum management tool.

FIG. 28 is a block diagram of one embodiment of a resource brokerage application.

FIG. 29 illustrates another example of a system diagram including an automated semantic engine and translator.

FIG. 30 illustrates a flow diagram of a method to obtain actionable data based on customer goals.

FIG. 31 illustrates a flow diagram of a method of implementation of actionable data and knowledge decision gates from total signal flow.

FIG. 32 illustrates a flow diagram of a method to identify knowledge decision gates based on operational knowledge.

FIG. 33 illustrates an overview of one example of information used to provide knowledge.

FIG. 34 is a map showing locations of three macrosites, 3 SigBASE units, and a plurality of locations evaluated for alternate or additional site deployment for a first example.

FIG. 35 is a graph of distribution of users by average downlink Physical Resource Block (PRB) allocation for the first example.

FIG. 36 illustrates rate of overutilization events and degree of overutilization for the first example.

FIG. 37A is a sector coverage map for three macrosites for the first example.

FIG. 37B illustrates signal strength for the sector shown in FIG. 37A for the first example.

FIG. 37C illustrates subscriber density for the sector shown in FIG. 37A for the first example.

FIG. 37D illustrates carrier-to-interference ratio for the sector shown in FIG. 37A for the first example.

FIG. 38A illustrates the baseline scenario shown in FIG. 34 for the first example.

FIG. 38B is a map showing locations of the three original macrosites and two additional macrosites for the first example.

FIG. 39 illustrates signal strength of the baseline scenario from FIG. 38A on the left and the scenario with two additional macrosites from FIG. 38B on the right for the first example.

FIG. 40A illustrates carrier-to-interference ratio of the baseline scenario from FIG. 38A for the first example.

FIG. 40B illustrates carrier-to-interference ratio of the scenario with two additional macrosites for the first example.

FIG. 41 illustrates a baseline scenario for a second example on the left and a map showing locations of the original macrosites from the baseline scenario with three additional proposed macrosites for the second example on the right.

FIG. 42 illustrates signal strength of the baseline scenario from FIG. 41 on the left and the scenario with three additional proposed macrosites from FIG. 41 on the right for the second example.

FIG. 43 illustrates carrier-to-interference ratio of the baseline scenario from FIG. 41 for the second example on the left and carrier-to-interference ratio of the scenario with three additional proposed macrosites from FIG. 41 on the right for the second example.

FIG. 44 illustrates a signal strength comparison of a first carrier (“Carrier 1”) with a second carrier (“Carrier 2”) for 700 MHz for a third example.

FIG. 45 illustrates carrier-to-interference ratio for Carrier 1 and Carrier 2 for the third example.

FIG. 46 is a graph of Area vs. RSSI and Traffic vs. RSSI for Carrier 1 and Carrier 2 for the third example.

FIG. 47 is a graph of traffic difference for Carrier 1 versus Carrier 2 for the third example.

FIG. 48 is a graph of SNR vs. RSRP for each SigBASE for the third example.

FIG. 49 is another graph of SNR vs. RSRP for each SigBASE for the third example.

FIG. 50 is a clustered graph of SNR vs. RSRP for each SigBASE for the third example.

FIG. 51 is another clustered graph of SNR vs. RSRP for each SigBASE for the third example.

FIG. 52 is a schematic diagram of a system of the present invention.

FIG. 53 illustrates one embodiment of a flow diagram of a channelizer configuration.

FIG. 54 illustrates another embodiment of a flow diagram of a channelizer.

FIG. 55 illustrates yet another embodiment of a flow diagram of a channelizer configuration.

FIG. 56A illustrates one embodiment of probability density functions per bin for noise and a signal with the noise.

FIG. 56B illustrates an example of a plurality of frequency bins and a critical value or threshold.

FIG. 56C illustrates another example of a plurality of frequency bins and a critical value or threshold.

FIG. 57A illustrates one example of probability on the channel level.

FIG. 57B illustrates one example of probability on the bin level.

FIG. 58A illustrates an example showing a critical value (γ), power levels of a noise signal, and power levels of a signal with noise.

FIG. 58B illustrates an example of the probability of false alarm for a noise signal.

FIG. 58C illustrates an example of the probability of missed detection for a signal with noise.

FIG. 59 illustrates one example of probabilities.

FIG. 60 illustrates an example of equations for hypothesis testing with channels for the following scenarios: (1) Noise Only (No Signal), Assert Noise Only, (2) Noise Only (No Signal), Assert Noise and Signal, (3) Noise and Signal, Assert Noise Only, and (4) Noise and Signal, Assert Noise and Signal.

FIG. 61 illustrates one example of an algorithm used in a channelizer.

FIG. 62A illustrates one example of a simulated fast Fourier transform (FFT), noise floor estimation, and comparison with channelization vectors.

FIG. 62B illustrates channelization vectors and comparisons for the FFT frame shown in FIG. 62A.

FIG. 62C illustrates amplitude probability distribution for the FFT frame shown in FIG. 62A.

FIG. 62D illustrates FFT frames grouped by block for the FFT frame shown in FIG. 62A.

FIG. 62E illustrates average power FFT by block for the FFT frame shown in FIG. 62A.

FIG. 62F illustrates the channelization vectors for the FFT frame shown in FIG. 62A. The x-axis indicates the frequency of the bin.

FIG. 62G illustrates the comparison vectors for the FFT frame shown in FIG. 62A.

FIG. 63A illustrates one embodiment of preemption by wideband detection.

FIG. 63B illustrates the example shown in FIG. 63A accounting for wideband detection.

FIG. 64 illustrates one embodiment of maximum resolution bandwidth (RBW) for vector resolution.

FIG. 65A illustrates one example of a spectrum scenario.

FIG. 65B illustrates an embodiment of channelization vectors.

FIG. 66A illustrates one example of bandwidth selection.

FIG. 66B illustrates an example using the bandwidth selection in FIG. 66A.

FIG. 67A illustrates another example of an FFT frame illustrating the center frequency on the x-axis and the frequency bin mean power level (Fbpower_mean) on the y-axis.

FIG. 67B illustrates channelization vectors and comparisons for the FFT frame shown in FIG. 67A.

FIG. 67C illustrates a graph for the FFT frame shown in FIG. 67A.

FIG. 67D illustrates a table for the FFT frame shown in FIG. 67A.

FIG. 68A illustrates an example of spectrum by channel.

FIG. 68B illustrates an example of channel detection for the spectrum shown in FIG. 68A.

FIG. 68C illustrates an example of a cascade for the spectrum shown in FIG. 68A.

FIG. 69A illustrates an example of a graph of total signal power.

FIG. 69B illustrates an example of a graph of total noise power.

FIG. 70A illustrates another example of a simulated FFT, noise floor estimation, and comparison with channelization vectors.

FIG. 70B illustrates channelization vectors and comparisons for the FFT frame shown in FIG. 70A.

FIG. 70C illustrates channelization vectors and comparisons for the FFT frame shown in FIG. 70A.

FIG. 71 illustrates one embodiment of a system including an antenna subsystem, an RF conditioning subsystem, and at least one front end receiver.

FIG. 72 illustrates one embodiment of a system including a channelizer (e.g., a frequency domain programmable channelizer), a blind detection engine, a noise floor estimator, and an I/Q buffer.

FIG. 73 illustrates one embodiment of a system including at least one data analysis engine.

FIG. 74 illustrates one embodiment of a system including an FFT engine.

FIG. 75 illustrates one embodiment of a system including the systems from FIGS. 71-73.

FIG. 76 illustrates one embodiment of a system including the systems from FIGS. 71-72 and 74.

FIG. 77 illustrates a flowchart of a channelizer using machine learning (ML) image comparison.

FIG. 78 illustrates a flowchart for the image used in the channelizer image comparison.

FIG. 79 illustrates a flowchart of a channelizer hypothesis test using machine learning (ML) image comparison.

FIG. 80 illustrates images used in the channelizer hypothesis test image comparison.

FIG. 81 illustrates a general purpose processing unit (GPPU) and central processing unit (CPU) architecture.

FIG. 82 represents a flow chart of data flow for the radio frequency general purpose processing unit (RFGPPU).

FIG. 83 illustrates a radio frequency general purpose processing unit (RFGPPU) and central processing unit (CPU) architecture.

DETAILED DESCRIPTION

The present invention is generally directed to spectrum analysis and management for electromagnetic signals, and more particularly for providing dynamic, prioritized spectrum utilization management.

In one embodiment, the present invention provides a system for autonomous spectrum management in an electromagnetic environment including at least one monitoring sensor operable to autonomously monitor the electromagnetic environment and create measured data based on the electromagnetic environment, at least one data analysis engine for analyzing the measured data, a semantic engine including a programmable rules and policy editor, and a tip and cue server, wherein the at least one data analysis engine includes a detection engine and a learning engine, wherein the detection engine is operable to automatically detect at least one signal of interest, wherein the learning engine is operable to learn the electromagnetic environment, wherein the programmable rules and policy editor includes at least one rule and/or at least one policy, wherein the tip and cue server is operable to use analyzed data from the at least one data analysis engine to create actionable data, and wherein the tip and cue server and the at least one data analysis engine are operable to run autonomously without requiring user interaction and/or input. In one embodiment, the tip and cue server is operable to activate an alarm and/or provide at least one report based on the actionable data. In one embodiment, the alarm is stored in a database for visualization. In one embodiment, the at least one monitoring sensor includes at least one antenna, at least one antenna array, at least one radio server, and/or at least one software defined radio. In one embodiment, one or more of the at least one monitoring sensor is integrated with at least one camera to capture video and/or still images. In one embodiment, the at least one data analysis engine further includes an identification engine, a classification engine, and/or a geolocation engine. In one embodiment, the system further includes a resource brokerage application, wherein the resource brokerage application is operable to optimize resources to improve performance of at least one customer application and/or at least one customer device. In one embodiment, the system further includes a certification and compliance application, wherein the certification and compliance application is operable to determine if at least one customer application and/or at least one customer device is behaving according to the at least one rule and/or the at least one policy. In one embodiment, the system further includes a survey occupancy application, wherein the survey occupancy application is operable to determine occupancy in frequency bands and schedule occupancy in at least one frequency band. In one embodiment, the actionable data indicates that the at least one signal of interest is behaving like a drone. In one embodiment, the semantic engine is operable to receive queries, searches, and/or search-related functions using natural language. In one embodiment, the semantic engine is operable to receive data as audio data, text data, video data, and/or image data. In one embodiment, the semantic engine is operable to create a semantic map including target data, wherein the semantic engine is operable to analyze related data and/or data with similar characteristics to the target data. In one embodiment, the system further includes a translator, wherein the translator is operable to receive data input including at least one use case, at least one objective, and/or at least one signal. In one embodiment, the learning engine is operable to use machine learning (ML), artificial intelligence (AI), deep learning (DL), neural networks (NNs), artificial neural networks (ANNs), support vector machines (SVMs), Markov decision process (MDP), natural language processing (NLP), control theory, and/or statistical learning techniques. In one embodiment, the learning engine is operable to compute a set of possible conditional probabilities depicting a set of all possible outputs based on input measurements to provide a predicted outcome using a data model. In one embodiment, the learning engine is operable to use third party data, wherein the third party data includes social media, population, real estate, traffic, geographic information system (GIS), network, signal site, site issue, and/or crowdsourced information. In one embodiment, the learning engine is operable to determine whether a data set processed and/or analyzed represents a sufficient statistical data set. In one embodiment, the learning engine includes a learning engine software development kit (SDK) operable to manage system resources relating to monitoring, logging, and/or organizing learning aspects of the system. In one embodiment, the semantic engine further includes a language dictionary. In one embodiment, the actionable data indicates that one or more of the at least one signal of interest is behaving like a drone. In one embodiment, the actionable data indicates a location for at least one macrosite and/or at least one tower. In one embodiment, the actionable data is created in near-real time.

In another embodiment, the present invention provides a system for autonomous spectrum management in a radio frequency (RF) environment including at least one monitoring sensor operable to autonomously monitor the RF environment and create measured data based on the RF environment, at least one data analysis engine for analyzing the measured data, a semantic engine including a programmable rules and policy editor, and a tip and cue server, wherein the at least one data analysis engine includes a detection engine, a geolocation engine, and a learning engine, wherein the detection engine is operable to automatically detect at least one signal of interest, wherein the learning engine is operable to learn the RF environment, wherein the programmable rules and policy editor includes at least one rule and/or at least one policy, wherein the semantic engine is operable to create a semantic map including target data, wherein the tip and cue server is operable to use analyzed data from the at least one data analysis engine to create actionable data, and wherein the tip and cue server and the at least one data analysis engine are operable to run autonomously without requiring user interaction and/or input. In one embodiment, the system further includes a visualization of the semantic map.

In yet another embodiment, the present invention provides a method for autonomous spectrum management in an electromagnetic environment including providing a semantic engine including a programmable rules and policy editor, wherein the programmable rules and policy editor includes at least one rule and/or at least one policy, monitoring the electromagnetic environment using at least one monitoring sensor and creating measured data based on the electromagnetic environment, analyzing the measured data using at least one data analysis engine, thereby creating analyzed data, wherein the at least one data analysis engine includes a detection engine and a learning engine, learning the electromagnetic environment using the learning engine, automatically detecting at least one signal of interest using the detection engine, and creating actionable data using a tip and cue server based on the analyzed data from the at least one data analysis engine, wherein the tip and cue server and the at least one data analysis engine are operable to run autonomously without requiring user interaction and/or input. In one embodiment, the method further includes the tip and cue server activating an alarm and/or providing at least one report based on the actionable data. In one embodiment, the method further includes the semantic engine creating a semantic map including target data. In one embodiment, the actionable data is created in near-real time.

Traditional management of spectrum is static, based on licenses that are geographical and band specific. The Federal Communications Commission (FCC) has allocated spectrum into a table. Utilization is increased by slicing the spectrum into finer slices. Additionally, interference is limited by imposing penalties by strict geographical band utilization rules and licenses. However, these traditional methods of spectrum management do not work with increasing demand and new services coming out. The new services would have to be at higher frequencies (e.g., above 10 GHz), which is very expensive and requires costly transceiver with a limited distance range.

Spectrum is valuable because it is a finite resource. Further, the demand for spectrum is ever-increasing. The Shannon-Hartley theorem calculates the maximum rate at which information can be transmitted over a communications channel of a specified bandwidth in the presence of noise as follows:

C
=

B
⁢
W
⁢

log
2

⁢

(

1
+

S
⁢
N
⁢
R

)

where C is the channel capacity in bits per second, BW is the bandwidth of the channel in Hz, and SNR is the signal-to-noise ratio.

Early attempts at managing spectrum include developing technology that increases spectrum efficiency (i.e., maximizing SNR). Although this results in more bits per Hz, the logarithmic function limits the gains in channel capacity resulting from improving technology. Additional attempts at managing spectrum also include developing technology to enable use of alternate spectrum (e.g., free-space optical (FSO) communication). However, using alternate spectrum, such as higher frequencies, leads to smaller ranges, line of sight limitations, increased elevation of transmission structures, and/or expensive infrastructure.

The missing component to spectrum management is bandwidth management. Bandwidth management provides flexible utilization of the spectrum, enables management of spectrum resources and users, while allowing spectrum usage to be quantified. The majority of applications using the spectrum can coexist if each application knows about the spectrum needs of other applications and how they plan to use the spectrum. However, because the needs of each application are dynamic, a dynamic spectrum management system is needed. The present invention allows autonomous, dynamic sharing of the electromagnetic spectrum to allow maximum utilization by diverse applications according to specific utilization rules (dynamic and/or static) while maintaining minimum interference between applications. This requires new tools that provide dynamic environmental spectral awareness of all signals present in the electromagnetic (e.g., radio frequency (RF)) environment to properly execute utilization rules, which are operable to describe or facilitate sharing spectrum resources among several competing users or protect one service user from others, among others.

5G requires spectrum awareness. Larger blocks of spectrum are required to support higher speeds. Dynamic spectrum sharing is necessary to make the spectrum assets available. Further, visibility of spectrum activity is required to support reliability targets. Interference avoidance and resolution must be embedded. Internet of Things (IoT)/machine communication wireless dependency elevates the need for real-time RF visibility to avoid disruption and safety concerns.

The system of the present invention provides scalable processing capabilities at the edge. Edge processing is fast and reliable with low latency. Environmental sensing processes optimize collection and analytics, making data sets manageable. Advantageously, the system minimizes backhaul requirements, allowing for actionable data to be delivered faster and more efficiently.

Deep learning techniques extract and deliver knowledge from large data sets in near-real time. These deep learning techniques are critical for identifying and classifying signals. Edge analytics further allow third party data (e.g., social media, population information, real estate information, traffic information, geographic information system) to further enrich captured data sets. A semantic engine and inference reasoner leverages insights generated by machine learning and edge analytics. Ontologies are established allowing for the creation of knowledge operable to inform and direct actions and/or decisions.

Referring now to the drawings in general, the illustrations are for the purpose of describing one or more preferred embodiments of the invention and are not intended to limit the invention thereto.

The present invention provides systems, methods, and apparatuses for spectrum analysis and management by identifying, classifying, and cataloging at least one or a multiplicity of signals of interest based on electromagnetic spectrum measurements (e.g., radiofrequency spectrum measurements), location, and other measurements. The present invention uses real-time and/or near real-time processing of signals (e.g., parallel processing) and corresponding signal parameters and/or characteristics in the context of historical, static, and/or statistical data for a given spectrum, and more particularly, all using baseline data and changes in state for compressed data to enable near real-time analytics and results for individual monitoring sensors and for aggregated monitoring sensors for making unique comparisons of data.

T The systems, methods, and apparatuses according to the present invention preferably are operable to detect in near real time, and more preferably to detect, sense, measure, and/or analyze in near real time, and more preferably to perform any near real time operations within about 1 second or less. In one embodiment, near real time is defined as computations completed before data marking an event change. For example, if an event happens every second, near real time is completing computations in less than one second. Advantageously, the present invention and its real time functionality described herein uniquely provide and enable the system to compare acquired spectrum data to historical data, to update data and/or information, and/or to provide more data and/or information on open space. In one embodiment, information (e.g., open space) is provided on an apparatus unit or a device that is occupying the open space. In another embodiment, the system compares data acquired with historically scanned (e.g., 15 min to 30 days) data and/or or historical database information in near-real time. Also, the data from each monitoring sensor, apparatus unit, or device and/or aggregated data from more than one monitoring sensor, apparatus unit, and/or device are communicated via a network to at least one server computer and stored on a database in a virtualized or cloud-based computing system, and the data is available for secure, remote access via the network from distributed remote devices having software applications (apps) operable thereon, for example by web access (mobile app) or computer access (desktop app). The at least one server computer is operable to analyze the data and/or the aggregated data.

The system is operable to monitor the electromagnetic (e.g., RF) environment via at least one monitoring sensor. The system is then operable to analyze data acquired from the at least one monitoring sensor to detect, classify, and/or identify at least one signal in the electromagnetic environment. The system is operable to learn the electromagnetic environment, which allows the system to extract environmental awareness. In a preferred embodiment, the system extracts environmental awareness by including customer goals. The environmental awareness is combined with the customer goals, customer defined policies, and/or rules (e.g., customer defined rules, government defined rules) to extract actionable information to help the customer optimize performance according to the customer goals. The actionable information is combined and correlated with additional information sources to enhance customer knowledge and user experience through dynamic spectrum utilization and prediction models.

The systems, methods, and apparatuses of the various embodiments enable spectrum utilization management by identifying, classifying, and cataloging signals of interest based on electromagnetic (e.g., radio frequency) measurements. In one embodiment, signals and parameters of the signals are identified. In another embodiment, indications of available frequencies are presented to a user and/or user equipment. In yet another embodiment, protocols of signals are also identified. In a further embodiment, the modulation of signals, data types carried by the signals, and estimated signal origins are identified. Identification, classification, and cataloging signals of interest preferably occurs in real time or near-real time.

Embodiments are directed to a spectrum monitoring unit that is configurable to obtain spectrum data over a wide range of wireless communication protocols. Embodiments also provide for the ability to acquire data from and send data to database depositories that are used by a plurality of spectrum management customers and/or applications or services requiring spectrum resources.

In one embodiment, the system includes at least one spectrum monitoring unit. Each of the at least one spectrum monitoring unit includes at least one monitoring sensor that is preferably in network communication with a database system and spectrum management interface. In one embodiment, the at least one spectrum monitoring unit and/or the at least one monitoring sensor is portable. In a preferred embodiment, one or more of the at least one spectrum monitoring unit and/or the at least one monitoring sensor is a stationary installation. The at least one spectrum monitoring unit and/or the at least one monitoring sensor is operable to acquire different spectrum information including, but not limited to, frequency, bandwidth, signal power, time, and location of signal propagation, as well as modulation type and format. The at least one spectrum monitoring unit is preferably operable to provide signal identification, classification, and/or geo-location. Additionally, the at least one spectrum monitoring unit preferably includes a processor to allow the at least one spectrum monitoring unit to process spectrum power density data as received and/or to process raw In-Phase and Quadrature (I/Q) complex data. Alternatively, the at least one spectrum monitoring unit and/or the at least one monitoring sensor transmits the data to at least one data analysis engine for storage and/or processing. In a preferred embodiment, the transmission of the data is via a backhaul operation. The spectrum power density data and/or the raw I/Q complex data are operable to be used to further signal processing, signal identification, and data extraction.

The system preferably is operable to manage and prioritize spectrum utilization based on five factors: frequency, time, spatial, signal space, and application goals.

The frequency range is preferably as large as possible. In one embodiment, the system supports a frequency range between 1 MHz and 6 GHz. In another embodiment, the system supports a frequency range with a lower limit of 9 kHz. In yet another embodiment, the system supports a frequency range with a higher limit of 12.4 GHz. In another embodiment, the system supports a frequency range with a higher limit of 28 GHz or 36 GHz. Alternatively, the system supports a frequency range with a higher limit of 60 GHz. In still another embodiment, the system supports a frequency range with a higher limit of 100 GHz. The system preferably has an instantaneous processing bandwidth (IPBW) of 40 MHz, 80 MHz, 100 MHz, or 250 MHz per channel.

The time range is preferably as large as possible. In one embodiment, the number of samples per dwell time in a frequency band is calculated. In one example, the system provides a minimum coverage of 2 seconds. The number of samples per dwell in time in the frequency band is calculated as follows:

N
s

≥

(

I
⁢
P
⁢
B
⁢
W

)

⁢

(
2
)

/
channel

The storage required in a buffer is a minimum of 2 seconds per channel per dwell time, which is calculated as follows:

storage
=

(

I
⁢
P
⁢
B
⁢
W

)

⁢

(
2
)

⁢

(

2
⁢

Bytes

)

⁢

(
channels
)

/

(

dwell
⁢

time

)

Spatial processing is used to divide an area of coverage by a range of azimuth and elevation angles. The area of coverage is defined as an area under a certain azimuth and range. This is implemented by antenna arrays processing, steerable beamforming, array processing, and/or directional antennas. In one embodiment, the directional antennas include at least one steerable electrical or mechanical antenna. Alternatively, the directional antennas include an array of steerable antennas. More antennas require more signal processing. Advantageously, spatial processing allows for better separation of signals, reduction of noise and interference signals, geospatial separation, increasing signal processing gains, and provides a spatial component to signal identification. Further, this allows for simple integration of geolocation techniques, such as time difference of arrival (TDOA), angle of arrival (AOA), and/or frequency difference of arrival (FDOA). This also allows for implementation of a geolocation engine, which will be discussed in detail infra.

Each signal has inherent signal characteristics including, but not limited to a modulation type (e.g., frequency modulation (FM), amplitude modulation (AM), quadrature phase-shift keying (QPSK), quadrature amplitude modulation (QAM), binary phase-shift keying (BPSK), etc.), a protocol used (e.g., no protocol for analog signals, digital mobile radio (DMR), land mobile radio (LMR), Project 25 (P25), NXDN, cellular, long-term evolution (LTE), universal mobile telecommunications system (UMTS), 5G), an envelope behavior (e.g., bandwidth (BW), center frequency (Fc), symbol rate, data rate, constant envelope, peak power to average power ratio (PAR), cyclostationary properties), an interference index, and statistical properties (e.g., stationary, cyclostationary, higher moment decomposition, non-linear decomposition (e.g., Volterra series to cover non-linearities, learning basic model).

The application goals are dependent on the particular application used within the system. Examples of applications used in the system include, but are not limited to, traffic management, telemedicine, virtual reality, streaming video for entertainment, social media, autonomous and/or unmanned transportation, etc. Each application is operable to be prioritized within the system according to customer goals. For example, traffic management is a higher priority application than streaming video for entertainment.

As previously described, the system is operable to monitor the electromagnetic (e.g., RF) environment, analyze the electromagnetic environment, and extract environmental awareness of the electromagnetic environment. In a preferred embodiment, the system extracts the environmental awareness of the electromagnetic environment by including customer goals. In another embodiment, the system uses the environmental awareness with the customer goals and/or user defined policies and rules to extract actionable information to help the customer optimize the customer goals. The system combines and correlates other information sources with the extracted actionable information to enhance customer knowledge through dynamic spectrum utilization and prediction models.

FIG. 1 illustrates one embodiment of an RF awareness and analysis system. The system includes an RF awareness subsystem. The RF awareness subsystem includes, but is not limited to, an antenna subsystem, an RF conditioning subsystem, at least one front end receiver, a programmable channelizer, a blind detection engine, a blind classification engine, an envelope feature extraction module, a demodulation bank, an automatic gain control (AGC) double loop subsystem, a signal identification engine, a feature extraction engine, a learning engine, a geolocation engine, a data analysis engine, and/or a database storing information related to at least one signal (e.g., metadata, timestamps, power measurements, frequencies, etc.). The system further includes an alarm system, a visualization subsystem, a knowledge engine, an operational semantic engine, a customer optimization module, a database of customer goals and operational knowledge, and/or a database of actionable data and decisions.

The antenna subsystem monitors the electromagnetic (e.g., RF) environment to produce monitoring data. The monitoring data is then processed through the RF conditioning subsystem before being processed through the front end receivers. The AGC double loop subsystem is operable to perform AGC adjustment. Data is converted from analog to digital by the front end receivers.

The digital data is then sent through the programmable channelizer, and undergoes I,Q buffering and masking. A fast Fourier transform (FFT) is performed and the blind detection engine performs blind detection. Additionally, the blind classification engine performs blind classification. Information (e.g., observed channels) is shared from the blind detection engine to the blind classification and/or the programmable channelizer (e.g., to inform logic and selection processes). Information from the blind detection engine is also sent to the envelope feature extraction module. Information from the blind classification engine is sent to the demodulation bank.

Information from the envelope feature extraction module, the demodulation bank, and/or the blind classification engine are operable to be used by the signal identification engine, the feature extraction engine, the learning engine, and/or the geolocation engine. Information from the AGC double loop subsystem, the I,Q buffer, masking, the programmable channelizer, the signal identification engine, the feature extraction engine, the learning engine, and the geolocation engine, the envelope feature extraction module, the demodulation bank, and/or the blind classification engine is operable to be stored in the database storing information related to the at least one signal (e.g., signal data, metadata, timestamps).

Information from the database (i.e., the database storing information related to the at least one signal), the signal identification engine, the feature extraction engine, the learning engine, and/or the geolocation engine is operable to be sent to the data analysis engine for further processing.

The alarm system includes information from the database storing information related to the at least one signal and/or the database of customer goals and operational knowledge. Alarms are sent from the alarm system to the visualization subsystem. In a preferred embodiment, the visualization subsystem customizes a graphical user interface (GUI) for each customer. The visualization system is operable to display information from the database of actionable data and decisions. In one embodiment, the alarms are sent via text message and/or electronic mail. In one embodiment, the alarms are sent to at least one internet protocol (IP) address.

The database of customer goals and operational knowledge is also operable to send information to a semantic engine (e.g., customer alarm conditions and goals) and/or an operational semantic engine (e.g., customer operational knowledge). The semantic engine translates information into constraints and sends the constraints to the customer optimization module, which also receives information (e.g., signal metadata) from the data analysis engine. The customer optimization module is operable to send actionable data related to the electromagnetic environment to the operational semantic engine. The customer optimization module is operable to discern which information (e.g., environmental information) has the largest statistically sufficient impact related to the customer goals and operation.

In one embodiment, the system includes at least one monitoring sensor, at least one data analysis engine, at least one application, a semantic engine, a programmable rules and policy editor, a tip and cue server, and/or a control panel as shown in FIG. 2.

The at least one monitoring sensor includes at least one radio server and/or at least one antenna. The at least one antenna is a single antenna (e.g., uni-directional or directional) or an antenna array formed of multiple antennas resonating at different frequency bands and configured in a 1D (linear), 2D (planar), or 3D (area) antenna configuration. The at least one monitoring sensor is operable to scan the electromagnetic (e.g., RF) spectrum and measure properties of the electromagnetic spectrum, including, but not limited to, receiver I/Q data. The at least one monitoring unit is preferably operable to autonomously capture the electromagnetic spectrum with respect to frequency, time, and/or space. In one embodiment, the at least one monitoring sensor is operable to perform array processing.

In another embodiment, the at least one monitoring sensor is mobile. In one embodiment, the at least one monitoring sensor is mounted on a vehicle or a drone. Alternatively, the at least one monitoring sensor is fixed. In one embodiment, the at least one monitoring sensor is fixed in or on a street light and/or a traffic pole. In yet another embodiment, the at least one monitoring sensor is fixed on top of a building.

In one embodiment, the at least one monitoring sensor is integrated with at least one camera. In one embodiment, the at least one camera captures video and/or still images.

In another embodiment, the at least one monitoring sensor includes at least one monitoring unit. Examples of monitoring units include those disclosed in U.S. Pat. Nos. 10,122,479, 10,219,163, 10,231,206, 10,237,770, 10,244,504, 10,257,727, 10,257,728, 10,257,729, 10,271,233, 10,299,149, 10,498,951, and 10,529,241, and U.S. Publication Nos. 20190215201, 20190364533, and 20200066132, each of which is incorporated herein by reference in its entirety.

In a preferred embodiment, the system includes at least one data analysis engine to process data captured by the at least one monitoring sensor. An engine is a collection of functions and algorithms used to solve a class of problems. The system preferably includes a detection engine, a classification engine, an identification engine, a geo-location engine, a learning engine, and/or a statistical inference and machine learning engine. For example, the geolocation engine is a group of functions and geolocation algorithms that are used together to solve multiple geolocation problems.

The detection engine is preferably operable to detect at least one signal of interest in the electromagnetic (e.g., RF) environment. In a preferred embodiment, the detection engine is operable to automatically detect the at least one signal of interest. In one embodiment, the automatic signal detection process includes mask creation and environment analysis using masks. Mask creation is a process of elaborating a representation of the electromagnetic environment by analyzing a spectrum of signals over a certain period of time. A desired frequency range is used to create a mask, and FFT streaming data is also used in the mask creation process. A first derivative is calculated and used for identifying possible maximum power values. A second derivative is calculated and used to confirm the maximum power values. A moving average value is created as FFT data is received during a time period selected by the user for mask creation. For example, the time period is 10 seconds. The result is an FFT array with an average of the maximum power values, which is called a mask.

The classification engine is preferably operable to classify the at least one signal of interest. In one embodiment, the classification engine generates a query to a static database to classify the at least one signal of interest based on its components. For example, the information stored in static database is preferably used to determine spectral density, center frequency, bandwidth, baud rate, modulation type, protocol (e.g., global system for mobile (GSM), code-division multiple access (CDMA), orthogonal frequency-division multiplexing (OFDM), LTE, etc.), system or carrier using licensed spectrum, location of the signal source, and/or a timestamp of the at least one signal of interest. In an embodiment, the static database includes frequency information gathered from various sources including, but not limited to, the Federal Communication Commission, the International Telecommunication Union, and data from users. In one example, the static database is an SQL database. The data store is operable to be updated, downloaded or merged with other devices or with its main relational database. In one embodiment, software application programming interface (API) applications are included to allow database merging with third-party spectrum databases that are only operable to be accessed securely. In a preferred embodiment, the classification engine is operable to calculate second, third, and fourth order cumulants to classify modulation schemes along with other parameters, including center frequency, bandwidth, baud rate, etc.

The identification engine is preferably operable to identify a device or an emitter transmitting the at least one signal of interest. In one embodiment, the identification engine uses signal profiling and/or comparison with known database(s) and previously recorded profile(s) to identify the device or the emitter. In another embodiment, the identification engine states a level of confidence related to the identification of the device or the emitter.

The geolocation engine is preferably operable to identify a location from which the at least one signal of interest is emitted. In one embodiment, the geolocation engine uses statistical approximations to remove error causes from noise, timing and power measurements, multipath, and non-line of sight (NLOS) measurements. By way of example, the following methods are used for geolocation statistical approximations and variances: maximum likelihood (nearest neighbor or Kalman filter); least squares approximation; Bayesian filter if prior knowledge data is included; and the like. In another embodiment, time difference of arrival (TDOA) and frequency difference of arrival (FDOA) equations are derived to assist in solving inconsistencies in distance calculations. In still another embodiment, angle of arrival (AOA) is used to determine geolocation. In yet another embodiment, power distribution ratio versus azimuth measurements are used to determine geolocation. In a preferred embodiment, geolocation is performed using Angle of Arrival (AOA), Time Difference of Arrival (TDOA), Frequency Difference of Arrival (FDOA), and power distribution ratio measurements. Several methods or combinations of these methods are operable to be used with the present invention because geolocation is performed in different environments, including but not limited to indoor environments, outdoor environments, hybrid (stadium) environments, inner city environments, etc.

The learning engine is preferably operable to learn the electromagnetic environment. In one embodiment, the learning engine uses statistical learning techniques to observe and learn an electromagnetic environment over time and identify temporal features of the electromagnetic environment (e.g., signals) during a learning period. In a preferred embodiment, the learning engine is operable to learn information from the detection engine, the classification engine, the identification engine, and/or the geolocation engine. In one embodiment, the learning function of the system is operable to be enabled and disabled. When the learning engine is exposed to a stable electromagnetic environment and has learned what is normal in the electromagnetic environment, it will stop its learning process. In a preferred embodiment, the electromagnetic environment is periodically reevaluated. In one embodiment, the learning engine reevaluates and/or updates the electromagnetic environment at a predetermined timeframe. In another embodiment, the learning engine reevaluates and/or updates the electromagnetic environment is updated after a problem is detected.

The statistical inference and machine learning (ML) engine utilizes statistical learning techniques and/or control theory to learn the electromagnetic environment and make predictions about the electromagnetic environment.

The survey occupancy application is operable to determine occupancy in frequency bands. In another embodiment, the survey occupancy application is operable to schedule occupancy in a frequency band. The survey occupancy application is also used to preprocess at least two signals that exist in the same band based on interference between the at least two signals.

The resource brokerage application is operable to optimize resources to improve application performance. In a preferred embodiment, the resource brokerage application is operable to use processed data from the at least one monitoring sensor and/or additional information to determine environmental awareness (e.g., environmental situational awareness). The environmental awareness and/or capabilities of a device and/or a resource are used to determine policies and/or reasoning to optimize the device and/or the resource. The resource brokerage application is operable to control the device and/or the resource. Additionally, the resource brokerage application is operable to control the at least one monitoring sensor.

The certification and compliance application is operable to determine if applications and/or devices are behaving according to rules and/or policies (e.g., customer policies and/or rules, government rules). In another embodiment, the certification and compliance application is operable to determine if the applications and/or the devices are sharing frequency bands according to the rules and/or the policies. In yet another embodiment, the certification and compliance application is operable to determine if the applications and/or the devices are behaving according to non-interferences rules and/or policies.

The sharing application is operable to determine optimization of how applications and/or devices share the frequency bands. In a preferred embodiment, the sharing application uses a plurality of rules and/or policies (e.g., a plurality of customer rules and/or policies, government rules) to determine the optimization of how the applications and/or the devices share the frequency bands. Thus, the sharing application satisfies the plurality of rules and/or policies as defined by at least one customer and/or the government.

The statistical inference and prediction utilization application is operable to utilize predictive analytics techniques including, but not limited to, machine learning (ML), artificial intelligence (AI), neural networks (NNs), historical data, and/or data mining to make future predictions and/or models. The system is preferably operable to recommend and/or perform actions based on historical data, external data sources, ML, AI, NNs, and/or other learning techniques. The semantic engine is operable to receive data in forms including, but not limited to, audio data, text data, video data, and/or image data. In one embodiment, the semantic engine utilizes a set of system rules and/or a set of system policies. In another embodiment, the set of system rules and/or the set of system policies is created using a prior knowledge database. The semantic engine preferably includes an editor and a language dictionary.

The semantic engine preferably further includes a programmable rules and policy editor. The programmable rules and policy editor is operable to include at least one rule and/or at least one policy. In one embodiment, the at least one rule and/or the at least one policy is defined by at least one customer. Advantageously, this allows the at least one customer to dictate rules and policies related to customer objectives.

The system further includes a tip and cue server. The tip and cue server is operable utilize the environmental awareness from the data processed by the at least one data analysis engine in combination with additional information to create actionable data. In a preferred embodiment, the tip and cue server utilizes information from a specific rule set (e.g., customer defined rule set), further enhancing the optimization capabilities of the system. The specific rule set is translated into optimization objectives, including constraints associated with signal characteristics. In a preferred embodiment, the tip and cue server is operable to activate at least one alarm and/or provide at least one report. In another embodiment, the tip and cue server is operable to activate the at least one alarm and/or provide the at least one report according to the specific rule set.

Advantageously, the system is operable to run autonomously and continuously. The system learns from the environment, and, without operator intervention, is operable to detect anomalous signals that either were not there before, or have changed in power or bandwidth. Once detected, the system is operable to send alerts (e.g., by text or email) and begin high resolution spectrum capture, or I/Q capture of the signal of interest. Additionally, the system is operable to optimize and prioritize applications using the learning engine.

FIG. 3 is a flow diagram of the system according to one embodiment.

FIG. 4 illustrates the acquisition component of the system. The system includes an antenna subsystem including at least one antenna, an analog front-end conditioning system, a radio receiver front-end system, and a I/Q buffer. The system is operable to perform control functions including, but not limited to, controlling a radio server, conditioning the radio server, I/Q flow control and/or time stamping, and/or buffer management.

FIG. 5 illustrates one embodiment of an analog front end of the system. In one embodiment, electromagnetic waves are sent directly to a radio receiver front-end subsystem as shown in Path A. Alternatively, the electromagnetic waves are sent through an analog filter bank and amplifier/channel with a filter (SSS), an amplifier (e.g., variable gain amplifier), and an automatic gain controller as shown in Path B before reaching the radio receiver front-end subsystem. In one embodiment, the BCU is 80 MHz. Alternatively, the BCU is 150 MHz. The radio receiver front-end subsystem is described in FIG. 6.

FIG. 6 illustrates one embodiment of a radio receiver front-end subsystem. Path A and Path B continue into a radio-frequency integrated circuit (RFIC), and then proceed to a digital down-converter (DDC) before downsampling (e.g., decimation) and moving through a field programmable gate array (FPGA). In one embodiment, signals from the FPGA are operable to be sent to a digital to analog converter (DAC). Alternatively, signals are sent via bus to a Universal Software Radio Peripheral hardware driver (UHD) host and SD controller before continuing to Path E, which is described in FIG. 7.

FIG. 7 continues the embodiment of the radio receiver front-end shown in FIG. 6 after digitization. In one embodiment, Path E continues to the I,Q buffer. In another embodiment, Path E continues to a baseband receiver. In one embodiment, signals are further processed using signal processing software (e.g., GNU Radio software). In yet another embodiment, the baseband receiver is connected to inputs and/or outputs. In one embodiment, the inputs include, but are not limited to, MicroSD Flash memory and/or a Universal Serial Bus (USB) console. In one embodiment, the outputs include, but are not limited to, USB 2.0 host and/or audio. Alternatively, data from the baseband receiver is sent to the I,Q buffer via the IGbE port.

The system preferably uses multiple receiver channels for the front end. In one embodiment, there are 4 receiver channels. Alternatively, there are 8, 12, 16, or 32 receiver channels. I,Q data is preferably tagged by the receiver channel and receiver antenna (e.g., bandwidth, gain, etc.) and then stored in the I,Q buffer before analysis is completed.

Advantageously, the system is hardware agnostic. The system is operable to provide a suggestion for hardware for a particular frequency set. Additionally, the hardware agnostic nature of the system allows for established architecture to persist. The system is cost effective because it also allows for cheaper antennas to be used, as well as less expensive filters, because calibration can be done using the system rather than the antennas and/or filters, as well as post-ADC processing to rectify any performance loss. Because the system processes all signals present in the spectrum and their inter-relationships to extract environmental awareness, so the analog front end does not require elaborate filtering to avoid interference and provide optimum dynamic range. Additionally, the analog front end does not require optimal antennas for all frequency bands and ranges to obtain environmental awareness.

For a time domain programmable channelizer, all filters' impulse responses must be programmable and the number of filters must be programmable. Additionally, the channel bandwidth resolution must be programmable starting from a minimum bandwidth. The center frequency of each channel must also be programmable. Decimation is based on channel bandwidth and desired resolution. However, these requirements are difficult to implement for channels with variable bandwidth and center frequency. Wavelet filters can be used effectively if the center frequency and channel's bandwidth follow a tree structure (e.g., Harr and Deubauchi wavelets). FIG. 8 is an example of a time domain programmable channelizer.

In a preferred embodiment, the system includes a frequency domain programmable channelizer as shown in FIG. 9. The programmable channelizer includes buffer services, preprocessing of the FFT bin samples, bin selection, at least one band pass filter (BPF), an inverse fast Fourier transform (IFFT) function, decomposition, and/or frequency down conversion and phase correction to yield baseband I,Q for channels 1 through R. The IFFT function and decimation function are done to obtain each decomposed channel I,Q at the proper sampling rate. Advantageously, the frequency domain programmable channelizer is more computationally efficient than a time domain programmable channelizer because each filter is just a vector in the frequency domain and the filtering operation is just a vector multiplication, decomposing the input signal into multiple channels of differing bandwidths is parsing the vector representing the input signal frequency domain content into a subvector of different length.

FIG. 10 is another embodiment of a programmable channelizer. Data enters the filter and channels generators with channelization selector logic for a table lookup of filter coefficient and channelization vectors. The programmable channelizer includes a comparison at each channel, which provides anomalous detection using a mask with frequency and power, which is then sent to the learning engine and/or the alarm system (“A”). Data processed with the FFT is sent to the blind detection engine and/or for averaging processing (“B”). In one embodiment, average processing includes blind detection of channel bandwidths and center frequency and comparison to resulting frequency domain channelization. Data from the table lookup of filter coefficient and channelization vectors undergoes preprocessing with a mix circular rotator to produce D₁blocks of R₁points. A sum is taken of the D₁block, and an R₁point IFFT is taken to produce discor overlap samples OL₁. This process occurs (e.g., in parallel) for D₁blocks of R₁points through DR blocks of RR points to produce OL₁through OL_R, which are then sent to the classification engine (“C”). All data from the I,Q buffer is preferably stored in a buffered database (“D”). In one embodiment, the I,Q buffer is partitioned into N blocks with L oversamples. In one embodiment, the original sample rate is decimated by D_iwhere i is from 1 to R.

In one embodiment, the channelizer is operable to receive information from an FFT engine and/or a FFT configuration module. In one embodiment, the FFT engine and/or the FFT configuration module are incorporated into the channelizer. Alternatively, FFT engine and/or the FFT configuration module are separate from and not incorporated into the channelizer. In one embodiment, the FFT engine and/or the FFT configuration module includes at least one channel definition, at least one FFT configuration, at least one deference matrix, at least one detector configuration, and/or at least one channel detection.

In one embodiment, the FFT engine and/or the FFT configuration module is operable to provide information to other parts of the blind detection engine (e.g., the noise floor estimator, the blind detection engine, the statistical channel occupancy algorithm), the channelizer, the at least one data analysis engine (e.g., the classification engine, the geolocation engine, the identification engine, the learning engine), the semantic engine, and/or the optimization engine.

FIG. 53 illustrates one embodiment of a flow diagram of a channelizer configuration. Channel Definitions define the channels to detect. Channelization Vectors define the channels within the context of FFT. FFT Configuration configures FFT with sufficient resolution bandwidth (RBW) to resolve ambiguity among the Channel Definitions. A Deference Matrix identifies narrowband (NB) channels that are subsets of wideband (WB) channels to avoid false detection and/or to detect simultaneous channels. Detector Configuration sets acceptable probability of False Alarm and probability thresholds for detection. Channel Detection determines which channels are present subject to the Detector Configuration and the Deference Matrix.

In one embodiment, the FFT Configuration is operable to capture channels in their entirety. In one embodiment, the FFT Configuration is operable to capture channels with a minimum number of points to support detection probabilities. In one embodiment, the FFT Configuration is operable to resolve ambiguities between channels by employing a sufficiently small RBW.

In one embodiment, a channelization vector is operable to specify normalized power levels per FFT bin for at least one channel (e.g., each channel). In one embodiment, the channelization vector power levels are preferably normalized with respect to peak power in the channel's spectrum envelope.

In one embodiment, a deference matrix is operable to identify channels with bandwidth that falls within the bandwidth of other channels (i.e., channels with wider bands). In one embodiment, the channels defer to the channels with wider bands. In one embodiment, the detector is operable to evaluate narrower band channels before the detector evaluates wider band channels. In one embodiment, positive detection of wider band channels is operable to further constrain criteria that defines the positive detection of deferring narrower band channels.

In one embodiment, noise floor estimation is operable to define the mean and standard deviation of noise from the average power FFT (e.g., block). In one embodiment, statistics are operable to be on a channel span level, a block containing multiple channels, or a bin level. In one embodiment, the statistics are operable to be used by the channel detector to set criteria for asserting channel detection.

In one embodiment, detector configuration is operable to set a minimum acceptable probability of false alarm. In one embodiment, the detector configuration is operable to set a minimum acceptable probability of missed detection.

In one embodiment, channel detection is operable to evaluate presence of channels in order of increasing bandwidths (i.e., narrowband channels before wideband channels). In one embodiment, the channel detection is operable to perform a hypothesis test for each bin using Noise Floor Estimation and maximum probability of false alarm. In one embodiment, the hypothesis test includes a first hypothesis that a bin contains only noise (H₀) and/or a second hypothesis that a bin contains noise and signal (H_a). In one embodiment, the channel detection is operable to determine the probability that S≤s bins within a given channel's bandwidth reject the “Noise only” hypothesis. In one embodiment, the channel detection is operable to assert that a channel is detected if the probability is greater than p_min. In one embodiment, the channel detection is operable to adjust the hypothesis test for deferring narrowband channels if detection probability is greater than p. In one embodiment, the channel detection is operable to set a new mean and a new standard deviation.

FIG. 54 illustrates another embodiment of a flow diagram of a channelizer. The Detector Configuration is operable to set a probability of false alarm per bin (α_bin) and a minimum probability (prob_min). The noise modeler is operable to estimate a mean noise per bin (μ_N) and a standard deviation per bin (σ_N). In one embodiment, the noise modeler is operable to determine a channelization vector (CV_bin). The channelizer is operable to slice the FFT per channel and determine k_obsas the number of bins above the threshold in the channelization vector. The probability calculator is operable to estimate β_bin, determine the probability of missed detection (p_md), determine the probability of false alarm (p_faor p-value), and determine the probability of detection being correct. The detector is operable to determine that a signal is detected if the probability is greater or equal to the minimum probability. The detector is operable to determine that a signal is not detected if the probability is less than the minimum probability.

FIG. 55 illustrates yet another embodiment of a flow diagram of a channelizer configuration. The noise modeler is operable to estimate a mean noise per bin (μ_N) and a standard deviation per bin (σ_N). The Detector Configuration is operable to set a probability of false alarm level per bin (α_bin), a power threshold (T_pwr), and a minimum probability of being correct (prob_min). The channelizer is operable to slice the FFT per channel and determine k_obsas the number of bins above the threshold in the channelization vector. The channelizer is operable to determine a mean of the signal and noise per bin (μ_N+S) and a standard deviation of the signal and noise per bin (σ_N+S). The probability calculator is operable to estimate β_bin, determine the probability of missed detection, determine the p-value, and determine the probability of being correct. The detector is operable to determine that a signal is detected if the probability is greater or equal to the minimum probability. The detector is operable to determine that a signal is not detected if the probability is less than the minimum probability.

FIG. 56A illustrates one embodiment of probability density functions per bin for noise and a signal with the noise. FIG. 56A includes a mean of the noise (μ_H₀), a mean of the signal with the noise (μ_H_a), a standard deviation of the noise (σ_H₀), a standard deviation of the signal with the noise (σ_H_a), a probability of false detection (α), a probability of missed detection (β), and a carrier-to-noise ratio (CNR). A minimum CNR is determined that satisfies α and β given σ_H₀and σ_H_aAdditionally, given N bins of channel bandwidth (chbw), a number of bins that reject H₀to determine presence of a channel is determined.

In one embodiment, the probability of false detection (a) is calculated using the following equation:

Pr
⁢

{

μ
ˆ

=

μ
X

|

H
0

⋂

(

X
≥

C
⁢
V

)

}

=
α

In one embodiment, the probability of correctly rejecting the noise only hypothesis is equal to the probability of getting as many as s bins with power above the threshold if the noise only hypothesis is true. In one embodiment, the probability of correctly rejecting the noise only hypothesis is calculated using the following equation:

Pr
⁢

{

ABT
≤
abt

|

H
0

⋂
Nbins

=
nbins

}

=

(

nbins

abt

)

⁢

α
abt

(

1
-
α

)

nbins
-
abt

⁢

where

 

⁢

⁢

abt
=

∑

i
=
1

NBins

(

X
i

≥

C
⁢
V

)

In one embodiment, the CNR is determined as follows:

CNR
=

μ

H
a

μ

H
0

In another embodiment, the CNR is determined as follows:

CNR
=

(

μ

H
a

-
CV

)

(

CV
-

μ

H
0

)

In one embodiment, CV=[μ_H₀+ϕ(1−α)σ_H₀]. The above equation is operable to be substituted as follows:

CNR
=

(

μ

H
a

-

[

μ

H
0

+

ϕ
⁡
(

1
-
α

)

⁢

σ

H
0

]

)

(

[

μ

H
0

+

ϕ
⁡
(

1
-
α

)

⁢

σ

H
0

]

-

μ

H
0

)