如何在Matlab中从差分方程获取非因果向量及构造分子分母向量

Hey there! Since you already have a solid foundation on causal system rules (h[n] = 0 for n < 0) and know how to use filter/filtic in MATLAB, let's break down your two questions with practical steps and examples.

1. How to Identify Non-Causal Vectors from a Difference Equation in MATLAB #

First, a quick recap: a system is non-causal if its output at time n depends on future inputs (i.e., x[k] where k > n). Here's how to spot this and extract the related vectors:

Step-by-Step Process: #

1. Rearrange your difference equation to group all output terms (y-related) on the left, and input terms (x-related) on the right. For example, take this non-causal equation:

   y[n] = 2x[n+1] + 0.5x[n] - 0.3y[n-1]
   

   Rearranged to:

   y[n] + 0.3y[n-1] = 2x[n+1] + 0.5x[n]
   

2. Check input term shifts: Look for any x[n+i] where i > 0 (these are future inputs). In the example above, x[n+1] is a future input, and its coefficient 2 is part of the non-causal vector.
3. Extract non-causal vectors in MATLAB:
   • List out all input coefficients and their corresponding shifts relative to n. For the example, that's:

     b_coeffs = [2, 0.5];  % Coefficients for x[n+1], x[n]
     shifts = [1, 0];      % Shifts: +1 (future), 0 (current)
     

   • Filter to get only the non-causal parts (shifts > 0):

     non_causal_coeffs = b_coeffs(shifts > 0);
     non_causal_shifts = shifts(shifts > 0);
     

   Now non_causal_coeffs holds the coefficients that make the system non-causal, and non_causal_shifts tells you how far into the future the system looks.

2. How to Construct Numerator (b) and Denominator (a) Vectors for MATLAB #

MATLAB's filter function is designed for causal systems by default, which use this standard difference equation form:

a₀y[n] + a₁y[n-1] + ... + aₙy[n-N] = b₀x[n] + b₁x[n-1] + ... + bₘx[n-M]

Where a = [a₀, a₁, ..., aₙ] (denominator) and b = [b₀, b₁, ..., bₘ] (numerator). For non-causal systems, we need to adapt the equation to fit this framework.

Example for a Non-Causal System: #

Let's use the same equation: y[n] = 2x[n+1] + 0.5x[n] - 0.3y[n-1]

1. Shift variables to make it "causal-compatible":
   Let k = n + 1 (so n = k - 1). Substitute into the equation:

   y[k-1] = 2x[k] + 0.5x[k-1] - 0.3y[k-2]
   

   Rearrange to match the standard filter form:

   y[k-1] + 0.3y[k-2] = 2x[k] + 0.5x[k-1]
   

2. Build the b and a vectors:
   • The right-hand side (input terms) gives us b: coefficients for x[k] and x[k-1] → b = [2, 0.5]
   • The left-hand side (output terms) gives us a: coefficients for y[k-1] and y[k-2] → a = [1, 0.3]
3. Calculate the non-causal impulse response:
   To get the original system's impulse response h[n] (where x[n] = δ[n]), we can compute the output of this shifted system and adjust the indices:

   % Construct impulse input x (x[k] = delta[k], so x(1) = 1, rest 0)
   x = zeros(1, 10);
   x(1) = 1;
   
   % Compute output y[k] (this corresponds to h[k-1] in the original system)
   y = filter(b, a, x);
   
   % Adjust indices to get h[n] for n from -4 to 5
   n = -4:5;
   h = [0, y(1:end-1)];  % Shift right by 1: h[-1] = y[0] = 2, h[0] = y[1] = 0.5 - 0.3*2 = -0.1, etc.
   
   % Visualize the result
   stem(n, h);
   xlabel('n');
   ylabel('h[n]');
   title('Impulse Response of Non-Causal System');
   

   Notice that h[-1] = 2 (non-zero for n < 0), which confirms the system is non-causal.

内容的提问来源于stack exchange，提问作者Zero_Cool