The modulo-11 algorithm yields a near-100% guarantee that, if any two accounts differ in a single digit, their check digits will also be different. This gives us a minimum Hamming distance of , which allows us to detect any single-digit error.
Find two coding methodologies in daily life and describe each of them.
Make a nonsingular code but not uniquely decodable. Then make a uniquely decodable code but not prefix code.
Examples
Note that Instantaneous ⇒ Uniquely Decodable ⇒ Non-singular
is an (and therefore uniquely decodable and non-singular).
is since no two codewords are the same. However it is not uniquely decodable since and are indistinguishable (it follows that it is not an instantaneous code either).
is since and have the same codeword (hence not uniquely decodable and not an instantaneous code).
is : when you get a it is the start of a new symbol and the previous symbol is given by counting how many 's since the previous . It is not a instantaneous code since the first codeword is a prefix of all the others.
is not an instantaneous code since is a prefix of . Since all codewords have even length we can rewrite it as a -ary code . It is easy to see that this is uniquely decodable since only ever appears as the second half of . Thus if an is followed by anything other than , it represents a .
Coding gain
The limit of coding gain of some code is its gain compared to Shannon limit.
What is the limit of coding gain of rate one half code at BER ?↓
Dual space
Dual space of , is an -dimensional subspace.
Each vector in is linearly independent and is in .
Linear block code
Generative matrix
从 code words 中选 个向量,末尾正好为单位阵
Code word
Parity check matrix
满足
Detect error
Syndrome of encoded message , .
Let . The syndrome of is
Hamming code
Construct (the parity-check matrix of) a hamming code of length 7? How about length 8?
评论