Data processing: artificial intelligence – Fuzzy logic hardware – Digital fuzzy computer
Reexamination Certificate
2001-02-26
2003-02-25
Davis, George B. (Department: 2121)
Data processing: artificial intelligence
Fuzzy logic hardware
Digital fuzzy computer
C706S008000, C706S050000, C706S052000
Reexamination Certificate
active
06526394
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of Invention
This invention relates to controllers. Specifically, the present invention relates to fuzzy controllers used in target detection systems to obtain a desired probability of false alarms.
2. Description of the Related Art
Target detection systems are used in a variety of demanding applications including radar air traffic control systems, missile target tracking systems, and electro-optical target detection systems employed on aircraft and ground-based military vehicles. Such applications often require accurate target detection systems that produce minimal false detections.
A target detection system typically includes an electromagnetic energy sensor that receives electromagnetic signals such as optical signals and outputs electronic signals in response thereto. A processing circuit analyzes the electronic output signals to determine if a target is present in the field of view of the sensor.
The sensor is often a focal plane array of electromagnetic energy detectors such as charge-coupled devices (CCDs). Detectors in the array may have different performance characteristics that may change with a changing signal environment. Often, the detectors are initially calibrated by aiming the sensor at a dark, uniform region of space. Electrical offset values or gain coefficients are applied to the outputs of the detectors to equalize the outputs and thereby compensate for detector signal non-uniformities.
In a typical constant false alarm rate (CFAR) target detection system, the processing circuit includes a detector non-uniformity correction circuit for performing the calibration, a background estimation circuit, and a threshold circuit. The background estimation circuit determines an initial background value that is subtracted from the outputs of the detectors to enhance signal-to-noise ratio. The threshold circuit establishes a detection voltage threshold range for the detectors in the array. Typically, a single threshold range is established for all detectors in the array.
An ‘alarm’ occurs when the magnitude of a detector output signal is within the threshold range. By controlling the threshold range, the target detection system can control the probability of making a false detection. However, decreasing the probability of false detection may increase the likelihood that a target will go undetected.
Use of a single threshold range for all detectors in the array is inefficient, as the performance capabilities of individual detectors are often not maximized. For example, low performance detectors may raise the desired lower threshold of the threshold range. Due to the higher threshold, the capability of any high performance pixels to detect targets in noisy environments is not utilized.
Detector background estimation is often performed when the target detection system is initially activated and is disabled thereafter. In existing systems, if the background estimation circuit remains enabled, target information may corrupt the background estimates. The corrupted values may greatly reduce the target detection capability of the system. Accordingly, many existing target detection systems fail to account for variations in background that often occur during system operation. As a result, the ability of such target detection systems to accurately detect targets is compromised.
Hence, a need exists in the art for an accurate target detection system that accounts for varying detector background levels and changing signal environments during system operation.
SUMMARY OF THE INVENTION
The need in the art is addressed by the accurate target detection system of the present invention. In the illustrative embodiment, the inventive system is adapted for use with electro-optical systems and includes a first circuit for receiving electromagnetic signals and providing electrical signals in response thereto. A second circuit corrects background non-uniformities and/or noise in the first circuit based on the electrical signals and provides calibrated electrical signals in response thereto. A third circuit determines if a target signal is present within the calibrated electrical signals and provides a target detection signal in response thereto. A fourth circuit selectively activates or deactivates the second circuit based on the target detection signal.
In a specific embodiment, the first circuit is an array of electromagnetic energy detectors, each detector providing an electrical detector output signal. The second circuit includes a non-uniformity correction circuit for compensating for gain non-uniformities and noise non-uniformities in the electromagnetic energy detectors. The second circuit includes a detector gain term memory for storing detector gain compensation values. The detector gain compensation values are normalized by noise estimates unique to each of the detectors. The third circuit includes a signal enhancement circuit for increasing the signal-to-noise ratio of the calibrated electrical signals. The third circuit includes a noise estimation circuit for estimating noise in each of the detector output signals and providing noise estimates in response thereto. The noise estimation circuit further includes a noise estimator and a recursive background estimator. The third circuit further includes a subtractor for subtracting background contained in the noise estimates from the calibrated electrical signals and providing background subtracted signals in response thereto. The signal enhancement circuit includes a frame integrator circuit for summing frames of image data and providing summed frames in response thereto. Each frame of image data contains data corresponding to the background subtracted signals. The signal enhancement circuit further includes a filter bank that enhances the signal-to-noise ratio of the summed frames and provides a filtered signal in response thereto. The third circuit includes a first threshold circuit for comparing the filtered signal to a first threshold and a second threshold and providing a threshold exceedance signal if the filtered signal is between the first threshold and the second threshold.
In the illustrative embodiment, the third circuit further includes a digital signal processor running a fuzzy controller that acilitates the determination of the first threshold by providing a threshold multiplier value The first threshold is a function of the threshold multiplier and noise variables for the background subtracted signals. The noise variables include a mean noise offset, noise variance estimates from the background subtracted signals, and a noise statistic for accounting for moments greater than two in noise statistics of the background subtracted signals. The fourth circuit includes a second threshold circuit for comparing the background subtracted signals to a target detection threshold and providing the target detection signal in response thereto when the background subtracted signals exceed the target detection threshold. The target detection signal acts as an inhibit signal and is input to background and noise estimation circuits to disable the estimation circuits when the target signal is possibly present with in the background subtracted signals. The first threshold is a function of the noise variance estimates and a second threshold multiplier.
The novel design of the present invention is facilitated by use of the second threshold circuit to selectively inhibit background updates and noise estimation calculations for the outputs of detectors of the sensor that possibly represent target data. As a result, background estimation functions and noise estimation functions may continue to run during operation of the target detection system of the present invention. This allows the target detection system of the present invention to track any changes in signal environment and detector noise performance during system operation This greatly increases the ability of the target detection system of the present invention to detect targets in noisy environments. Furthermore, use of threshold levels unique to eac
Baker Todd L.
Caber Timothy E.
Dang Hien T.
Lo Thomas K.
Wald Sheldon S.
Alkov Leonard A.
Davis George B.
Lenzen, Jr. Glenn M.
Raufer Colin M.
Raytheon Company
LandOfFree
Accurate target detection system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Accurate target detection system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Accurate target detection system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3180775