Application of Data Compression to the MIL-STD-1553 Data Bus

1
Application of Data Compression to
theMIL-STD-1553 Data Bus

• Scholars Day
• Feb. 1, 2008
• By Bernard Lam

2
Overview

• Background
• MIL-STD-1553
• Bus Trace Analysis
• Solutions Compression Algorithms
• Zero-Tracking, Modified Run-Length, and
  Differential
• Error Analysis
• Conclusions Future Research

3
Goal Of Research

• To extend the bandwidth capabilities of
  MIL-STD-1553 Bus, using compression techniques.
• Develop algorithms suitable for legacy systems
• Demonstrate that the time to compress and
  decompress data is offset by the overall savings
  in data transmission time.

4
Timing Analysis
Timing Diagram
5
Background MIL-STD-1553

• MIL-STD-1553 serial data bus
• Developed in the late 1960s and early 1970s
• Limited/Low Bandwidth
• 1 Mb/s
• Has lead to development of multiple independent
  busses
• Time division multiple (TDM) access

System Model
6
Background MIL-STD-1553

• MIL-STD-1553 (contd)
• Manchester Bi-phase encoding
• Data word size 16 bit
• Sync Waveform
• Parity Bit

Message Format
7
Background MIL-STD-1553

• MIL-STD-1553 (contd)
• Max. single-command transmission size of 32 words
• Safety and Mission Critical System
• Real-Time System
• Replacement of MIL-STD-1553 with updated bus
  protocol, such as Fibre Channel, not a viable
  solution because of extensive costs.

8
Bus Trace Analysis

• Analysis was conducted using data from multiple
  bus traces of data captured at the F/A 18
  Advanced Weapons Laboratory.
• Each trace represented roughly 30 seconds of
  flight data and included examples of mode changes
  and start-up conditions.

9
Bus Trace Analysis

• Significant amount of zeros

Percent of Zeros
5 Hz
10 Hz
20 Hz
78.6
90.1
96.3
Max Zeros
72.0
88.5
53.5
Min Zeros
73.5
88.8
68
Avg. Zeros
10
Bus Trace Analysis

• Limited number of changes between
  consecutive message transmissions

Percent of Changes
5 Hz
10 Hz
20 Hz
78.6
27.5
21.7
Max Changes
2
0
2.0
Min Changes
3.3
3.3
3.9
Avg. Changes
11
Data Compression

• Lossless vs. Lossy Compression
• Lossless
• Original data is completely retrievable by means
  of decompression
• Ex. Winzip, GIF
• Lossy
• Lose information original data not retrievable
  when decompressed
• Higher Compression Ratios
• E.g., jpeg, mpeg, mp3

12
Data Compression

• Coding Performance and Efficiency
• Measured by compression ratio

13
Data Compression

• Criteria
• Lossless Compression
• Take advantage of message format of MIL-STD-1553
• Limit worst case expansion
• Limit computational and memory requirements

14
Compression Algorithms

• Common Value Tracking
• Zero-Tracking
• Modified Run-Length Encoding
• Differential Encoding

15
Zero Tracking

• Encodes long sequences containing mostly zeros
• Uses marker sequence to indicate the position of
  zeros
• Transmits
• Position Address (marker sequence)
• Non-Zero Data Words

16
Zero-Tracking Encoding (Example)
Encoded Data (Hex)
ZT
Input Data (Hex)
Word Count (Hex)
CBD0
0
0
1
FFFF
0
1
1
59
FFFF
2
0
AC9F
59
3
0
486
0
4
AC9F
5
0
6
0
7
1
0
8
0
9
1
486
A
0
0
B
1
17
Zero Tracking

• If a 32-word block is compressed
• 2 data words are required to indicate positions
• Can transmit maximum of 31 uncompressed data
  words
• Most significant bit in 1st address word is used
  to indicate if uncompress/compressed
• Worst Case Compression Ratio
• comp. ratio 31/32

18
Modified Run-Length Encoding

• Encodes consecutive sequences of identical words
• Uses marker sequence to indicate the presence of
  repeated sequences within block set
• For block of 32 words
• Worst Case Expansion 31/32

19
Modified Run-Length (Example)
Encoded Data (Hex)
RT
Input Data (Hex)
Word Count (Hex)
67A0
0
0
0
0
0
1
1
FFFF
0
2
1
5604
FFFF
3
0
0
9840
5604
4
1
B1F4
5604
5
1
5604
6
1
5604
7
1
5604
8
9840
9
0
9840
A
1
B1F4
B
0
20
Differential Encoding

• Encodes only changes of previous vs. current word
  locations
• A differential scheme takes advantage of the fact
  that for a given rate group one transmission to
  the next does not change
• Two buffers are required for comparison of
  previous and current transmissions

21
Differential Encoding
Encoded Data (Hex)
DT
Current Data (Hex)
Previous Data (Hex)
Word Count (Hex)
20D0
0
0054
0054
0
12F8
0
0815
0815
1
9FB2
1
12F8
AF58
2
FDA9
0
0000
0000
3
A14F
0
0000
0000
4
0
6542
6542
5
0
FFFF
FFFF
6
0
FFFF
FFFF
7
1
9FB2
2222
8
1
FDA9
8966
9
0
8966
8966
A
1
A14F
0052
B
22
Compression Ratios
Average Compression Ratios For Algorithms
5 Hz
10 Hz
20 Hz
MC2
MC1
MC2
MC1
MC2
MC1
2.60
2.44
3.39
4.65
1.66
2.63
Zero-Tracking
1.17
2.13
2.80
2.44
1.97
1.34
Mod. Run-Length
1.31
8.37
7.22
14.47
5.74
12.47
Differential
23
Compression Bit Status

• 1st Bit of 1st 16-bit word indicates the
  compression status
• 1 - equals uncompressed
• 0 equals compressed

Block Set Format
Compression Status
Bit Position Word
0 15 bits
16 bits 16 bits
30 Data Words 30 Data Words
1 15 bits
31 Data Words 31 Data Words
Bit Position Word
31 16 bit Data Words
30 16 bit Data Words
24
Transmission Error Effects

• Effects of data errors can be amplified when
  using data compression
• If higher levels of error detection and
  correction (EDAC) are needed, one or more data
  words can be dedicated to EDAC

25
Transmission Error Effects

• Standard 1553 Error Checking
• Bit Errors can be detected
• Exception multiple-bit errors without parity
  change cannot be detected
• Common Value Tracking
• If an undetected error is in the bit position
  word, multiple words can be corrupted.
• If an undetected error is in the data word, only
  that word location is impacted

26
Transmission Error Effects

• Modified Run-Length Compression
• Like zero tracking a error in the bit position
  word can invalidate a run
• Error dramatically worse result than that of
  zero-tracking
• Differential Encoding
• Error in address word can result incorrect
  updating
• Worst Case All data words are updated
• Further Research Required

27
Future Research

• Error Handling Routines
• Effects of mode-changing and start-up
• Timing analysis for Run-Length and Differential
  Encoding

28
Conclusions

• Reviewed Statistical Analysis of Trace Data
• Able to achieve compression ratios greater than
  one for all algorithms
• Discussed Error Analysis
• Preliminary timing simulations of timing look
  promising

29
Acknowledgements

• Dr. Russell Duren
• Dr. Michael Thompson

30
QUESTIONS?