Image analysis – Histogram processing – For setting a threshold
Patent
1988-09-27
1991-10-22
Razavi, Michael
Image analysis
Histogram processing
For setting a threshold
382 8, 382 1, 364513, G06K 962
Patent
active
050602799
DESCRIPTION:
BRIEF SUMMARY
TECHNICAL FIELD
This invention relates to expert systems. Such systems can be used for monitoring and fault diagnosis in a wide variety of applications.
PRIOR ART
Conventional expert systems are usually operated on computers and are used in such applications such as medical diagnosis, geophysical prospecting, electronic equipment configuration and planning of maintenance and control strategies. Such systems are usually rule based systems in which decision making follows a tree-like configuration in which a yes or no type decision is made at each branch along the tree. Fault diagnosis involves relating patterns of symptoms to specific faults or to corresponding corrective courses of action. The conventional expert system involves acquisition of information through extensive discussions with technicians or engineers familiar with the equipment or system to be diagnosed. This acquisition of information is an extremely time consuming and costly process. Furthermore it sometimes, involves a user predicting the implications of faults in a theoretical way and it is doubtful whether this is practical as it involves design staff in work which can approach the complexity of the original system design process. There is thus a need for an expert system which is much simpler and less time consuming to evolve. The present invention is directed to the provision of such a system.
SUMMARY OF THE INVENTION
According to the present invention there is provided an expert system which can be used for fault diagnosis and maintenance purposes which comprises processing means which receive data relating to equipment or system to be diagnosed, where the processing means is arranged to manipulate input data in a given manner to produce data representative of a parameter or parameters which can indicate the condition of the equipment or system, characterised in that said processing means includes an adaptive pattern recognition facility in the form of an adaptive combiner, which is capable of instruction to recognize particular data combinations as representative of particular conditions whereby after instruction said processing means operates on input data to provide an indication of the condition of equipment or systems under diagnosis.
The input data can be fed to the processing means from the equipment or system under test or can input via a keyboard.
The system can be instructed by connecting it to an equipment or system which is known to have essentially no faults. Data from that equipment is fed to the expert system and the weights of the adaptive combiner are adjusted according to a recursive means squares technique. A known fault is then introduced into the equipment and data again fed to the system. The weights are again adjusted to form a best fit to the data already introduced using the same recursive means squares error technique. This process is carried out for different degrees of the known fault and at each stage the processor is instructed as to the degree and type of fault. This process can be repeated for different faults to generate a series of combiners each having a set of weights updated to form a best fit to the input data.
Once instructed in this way the system can then be used to receive data from equipment to be diagnosed. The data to be processed is applied to the adaptive combiner which generates an output indicative of any fault existing in the equipment. The output or outputs can be used to drive a visual display unit which can provide a visible indication of the type and degree of fault. In addition, outputs can be generated which indicate the action required to correct the fault.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention will be described now by way of example only with particular reference to the accompanying drawings. In the drawings:
FIG. 1 is a block schematic diagram illustrating an expert system in accordance with the present invention;
FIG. 2 is a block schematic diagram of a digital radio system;
FIG. 3 shows a constellation type display for a digital radio system;
FIG. 4 is a
REFERENCES:
patent: 4376298 (1983-03-01), Sokol et al.
patent: 4449240 (1984-05-01), Yoshida
patent: 4517468 (1985-05-01), Kemper et al.
patent: 4635214 (1987-01-01), Kasai et al.
patent: 4703446 (1987-10-01), Momose
patent: 4739492 (1988-04-01), Cochran
Inspec. Abstract No. 86C010699, Insepc IEE (London) & IEE Coll. on "Adaptive Filters", Digest No. 76, Oct. 10, 1985, Crawford et al. Adaptive Pattern Recognition Applied To An Expert System For Fault Diagnosis In Telecommunications Equipment, pp. 10/1-8.
Inspec. Abstract No. 84C044315, Inspec IEE (London) & IEE Saraga Colloquium on Electronic Filters, May 21, 1984; Rutter et al., "The Timed Lattice--A New Approach To Fast Converging Equaliser Design", pp.VIII/1-5.
W. R. Simpson & C. S. Dowling, "WRAPLE: The Weighted Repair Assistance Program Learning Extension", IEEE Design & Test, vol. No. 2, Apr. 1986; pp. 66-73.
B. B. Dunning, "Self-Learning Data-Base For Automated Fault Localization", IEEE, 1979; pp. 155-157.
R. M. Stewart, "Expert Systems For Mechanical Fault Diagnosis", IEEE, 1985; pp. 295-300.
Cowan Colin F. N.
Crawford Thomas M.
Marton Virgil N.
Hewlett--Packard Company
Razavi Michael
LandOfFree
Expert system using pattern recognition techniques does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Expert system using pattern recognition techniques, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Expert system using pattern recognition techniques will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-115881