A GENERIC COMPONENT BASED EXPERT SYSTEM SHELL FOR AIRBORNE EQUIPMENT DESIGN

1
A GENERIC COMPONENT BASED EXPERT SYSTEM SHELL FOR
AIRBORNE EQUIPMENT DESIGN

• B.Ramesh Kumar1, J.Shanmugam1, S.Janarthanan2
  R.Santhiseela2
• 1 Madras Institute of Technology, India
• 2 Defence RD Organisation, India

2
ObjectiveExpert System for GAS

• Expert System to provide Guidelines
• Architecture Selection
• Environmental Testing
• EMI/EMC
• Reliability Engineering
• Testability
• Power
• Expert System for Auditing - To Provide Testing
  Procedure for the equipment component and help
  the designer with the evaluation process
• Expert System for Searching

3
Why such a System is needed?

• Problem 1 Interlinking of various domains
• The design of airborne equipment requires
  expertise knowledge in various interdependent
  domains
• Thus the simultaneous processing of all domains
  make the design complex
• If the airborne equipment is going to be critical
  in its function, then the design becomes more
  complex
• All these burdens the designer
• Problem 2 Voluminous of Parameters
• In the design of airborne equipment lots of
  parameters have to be considered in detail
• Due to voluminous of data and parameter, the
  designer can leave some parameters unnoticed or
  may skip one or more design steps

4
Expert System

• The British Computer Societys specialist group
  on Expert System produced the following formal
  definition An Expert System is regarded as the
  embodiment within a computer of a knowledge-based
  component, from an expert skill, in such a form
  that the system can offer intelligent advice or
  decision about a processing function

5
How KNOWLEDGE BASE for Airborne Equipment Design
is Framed?
Generalised
Example
Block_Name Value_1 Value_2
Value_n End
Architecture_Candidate Processor
Buses Topology End
Processor Type_Of_Processor
Speed Throughput Weight End
6
How INFERENCE ENGINE in Airborne Equipment Design
Expert System Works?
Architecture_Candidate Processor
Buses Topology End
Using Forward Chaining Method
Processor Type_Of_Processor
Speed Throughput Weight End
Buses Bus_Width Bus_Speed
Transmission End
7
Generic Components Based Expert System Shell
Knowledge Base 1
Knowledge Base 2
Generic Inference Engine
User Interface
Knowledge Base 3
. . .
Knowledge Base n
8
Architecture Selection Process

• Frame Standard Metrics
• Determine Physical Constraints
• Select Architecture Candidates appropriately

• Architecture Candidates
• Computation Element
• Communication Element
• Configuration

9
User Interface ArchitectureGuidelines
10
User InterfaceReliability Engg.
11
User InterfaceEnvironmental Testing
12
User InterfaceTestability DesignGuidelines
13
Testability

• Design Guidelines
• Evaluation of Equipment
• Equipment Digital Circuits Analog Circuits
  PSU RF
• Evaluation of Digital Circuits

14
User InterfaceTestability Evaluation of Airborne
Equipment
15
Testability Evaluation of EquipmentSome areas in
Design some sample Questions
Design Areas Sample Questions Weight (0 ? Weight ? 10) Weight (0 ? Weight ? 10) Weight (0 ? Weight ? 10)
Design Areas Sample Questions Expert 1 Expert 2 Expert 3
Mechanical Design Is a standard grid layout used on boards to facilitate identification of components?   Is enough spacing provided between components to allow for clips and test probes? 3 6 6 8 4 5
Partitioning Is each function to be tested placed wholly upon one board?   If more than one function is placed on a board, can each be tested independently? 7.5 8 6 7 8 9
Parts Selection Is the number of different part types the minimum possible?   Have parts been selected which are well characterised in terms of failure modes? 9 8.5 8 9 8.5 8
Analog Design Is one test point per discrete active stage brought out to the connector?   Is each test point adequately buffered or isolated from the main signal path? 8.2 8.5 7 8 4 9
Digital Design Does the design contain only synchronous logic?   Are all clocks of differing phases and frequencies derived from a single master clock? 7.8 8 NA NA 9 8
16
Testability Evaluation of EquipmentStatistical
Analysis

• Some Experts expertise in particular field like
  Analog, Digital, PSU etc.,
• So more Weightage is given to the score, given by
  the Expert of that particular field

Analog Expert
Analog Design
BIT
More Weightage
Digital Design
PSU
17
Testability Evaluation of Digital Circuits

• Testability, TY f (Controllability,Observability
  )
• Testability Measures - studied
• SCOAP (Sandia Controllability Observability
  Analysis Program)
• TMEAS (Testability MEASurement program)
• CAMLOT (Computer-Aided Measure for LOgic
  Testability)
• CAMLOT was chosen

18
Testability Evaluation of Digital Circuits ATPG
- Modified FAN (FANout algorithm)

• Propagate the fault to Primary Output (PO)
• Backtrace from PO to all Primary Inputs (PIs)
• Proceed with forward tracing from PIs to all Pos
• ATPG Algorithms analysed
• D-Algorithm
• PODEM
• FAN
• FAN Algorithm is chosen, and modified to suit our
  need

19
Testability Evaluation of Digital Circuits
Working of Modified FAN Algorithm
A B C
A B C
Y
Y
X
A B C
A B C
Y
Y
1. Fixing fault 2. Making Line to be fault 3.
Propagate the fault to (PO 4. Backtrace PO value
to PI 5. Find all Line values (Test patterns)
A B C
Y
20
User InterfaceTestability Evaluation
21
Electromagnetic Interference/ Compatibility(Appli
cable to Airborne Equipments Excluding RF)
22
User InterfaceElectromagnetic Interference/
Compatibility
23
User InterfaceEMI/ EMC Evaluation
2
1
3
24
User InterfaceElectric PowerGuidelines
25
User InterfaceExpert System Based Guidelines
Search

1. Enter the Question
2. On search, identifies the keywords and searches
   for them
3. It displays the matches found
4. It asks the user to select the preferred match
5. It displays the guidelines for the selected match

26
Organisation of Knowledge Base for Expert Search
Rule Structure Blockname Predecessor end
Example B1 B0.1 end B1.1 B1 end B1.3 B1 end B1.
3.1 B1.3 end
How Inference Engine Works Here? Searching for
B1.3.1(end branch of a tree) leads to the
identification of B1.3 which in turn identifies
B1. Similarly the iteration continues till it
finds the root (B0.1)
27
Who can use this System?

• Fresh Designer (as Study Material and as thumb
  rules for design)
• Designer (during Design process)
• Designer (after Design is complete for Evaluation)

28
References

• 1 Dutta.S, 1997, Strategies For Implementing
  Knowledge Based Systems, 20132, IEEE Trans.
  Engineering Management, pp. 79-90.
• 2 Santhiseela.R and Janarthanan.S, 2003, An
  Expert System For Automatic Fault Diagnosis Of A
  Quadruplex Digital Computer, International Conf
  on Advances in Aerospace Science, pp. 294-301.
• 3 Spitzer.R.Cary, 1993, Digital Avionics
  Systems Principles And Practices, Ed 2, MGH Inc.
• 4 James.P.Ignizio, 1991, Introduction To Expert
  Systems The Development And Implementation Of
  Rule Based Expert System, NY, MGH Inc.
• 5 Spitzer.R.Cary, 2001, The Avionics Handbook,
  NY, CRC Press.
• 6 Donald.A.Waterman, 1985, A Guide To Expert
  Systems, MA, Addison-Wesley Pubs Co.
• 7 James.N.Siddall, 1990, Expert System For
  Engineers, NY, Marcel Dekker Inc.
• 8 Dickman.T.J and Roberts.T.M, 1988, Modular
  Avionics System Architecture Decision Support
  System, IEEE 88CH2596-5, Proc. IEEE 1988 NAECON,
  pp.1549-1552.
• 9 MIL-HDBK-338B, 1998, Electronic Reliability
  Design Handbook, USA, DOD.

29
References

• 10 MIL-STD-810F, 2000, Test Method For
  Environmental Engineering Considerations And
  Laboratory Tests, USA, DOD.
• 11 MIL-STD-461D, 1993, Requirements For The
  Control Of Electromagnetic Interference Emission
  And Susceptibility, USA, DOD.
• 12 MIL-HDBK-1857, 1998, Grounding Bonding and
  Shielding Design Practices, USA, DOD.
• 13 MIL-STD-2165, 1985, Testability Program For
  Electronic Systems And Equipment, USA, DOD.
• 14 Kovijanic.P.G, 1979, Testability Analysis,
  IEEE Test Conference, Digest Of Papers,
  pp.310-316.
• 15 MIL-STD-704E, 1991, Aircraft Electric Power
  Characteristics, USA, DOD.
• 16 Bennetts.R.G, Maunder.C.M and Robinson.G.D,
  1981, CAMLOT A Computer Aided Measure Of Logic
  Testability, Vol. 2, Proc. IEEE International
  Conference On Circuit and Computers.
• 17 Fujiwara.H and Shimono.T, 1983, On The
  Acceleration Of Test Generation Algorithms, Vol.
  C-32, IEEE Trans. Computers, pp. 1137-1144.