Implementation of a (31, 16) BCH code on a Microcontroller
Error correcting codes are used throughout digital communication systems today. Cell-phones, CD players, satellites, digital pagers and many other communication devices all use varying amounts of error control to achieve a certain degree of accuracy in transmitting information. The idea behind error control codes is very simple; one can insert some amount of controlled redundancy into an information sequence before transmitting it through an analog channel where it will be exposed to noise while on its way to a destination (the receiving end). The extra redundancy will help the receiver (to some degree) detect and possibly correct errors that occurred during transmission.
In our project we focused on implementing a (31, 16) triple error correcting binary BCH code. A BCH code is a type of block code; this means that information is split into blocks of a specified length and each block is separately encoded and decoded independently of the other blocks. In our code the number 31 represents the length of the code which is expressed in bits, and the number 16 represents the number of information bits per block. Thus our (31, 16) code has a redundancy of 31 – 16 = 15 bits per 31 bit block, which is a little under 50%. In general, the more redundant a code is, the smaller the data rate of the code is, that is the more total data one has to send per information bit.
In this project we designed and implemented our (31, 16) BCH code to run on the Atmel Mega32 microcontroller. We interfaced with the Hyperterm program on the computer in order to test our code on real data. When starting the project, we found that it was easier to first program and test our code at home on a C compiler and then later modify the program to work under the Cvavr embedded environment in lab.
Link: http://instruct1.cit.cornell.edu/courses/e...bk26/index.html
Error correcting codes are used throughout digital communication systems today. Cell-phones, CD players, satellites, digital pagers and many other communication devices all use varying amounts of error control to achieve a certain degree of accuracy in transmitting information. The idea behind error control codes is very simple; one can insert some amount of controlled redundancy into an information sequence before transmitting it through an analog channel where it will be exposed to noise while on its way to a destination (the receiving end). The extra redundancy will help the receiver (to some degree) detect and possibly correct errors that occurred during transmission.
In our project we focused on implementing a (31, 16) triple error correcting binary BCH code. A BCH code is a type of block code; this means that information is split into blocks of a specified length and each block is separately encoded and decoded independently of the other blocks. In our code the number 31 represents the length of the code which is expressed in bits, and the number 16 represents the number of information bits per block. Thus our (31, 16) code has a redundancy of 31 – 16 = 15 bits per 31 bit block, which is a little under 50%. In general, the more redundant a code is, the smaller the data rate of the code is, that is the more total data one has to send per information bit.
In this project we designed and implemented our (31, 16) BCH code to run on the Atmel Mega32 microcontroller. We interfaced with the Hyperterm program on the computer in order to test our code on real data. When starting the project, we found that it was easier to first program and test our code at home on a C compiler and then later modify the program to work under the Cvavr embedded environment in lab.
Link: http://instruct1.cit.cornell.edu/courses/e...bk26/index.html