Göm meny

Errata

Signal Processing by Fredrik Gustafsson, Mille Millnert and Lennart Ljung

This errata focuses on factual errors that could interfere with the understanding of the presented material. General typos are usually not entered, non the less we still appreciate getting informed about them.

  • p13L5: n(t) should be n[k].
  • p15L7: h_k[l] should be h[k] .
  • p28L6: x(t)=0.5 should be x(t)=0.
  • p31: In the caption, the second 'e' should be 's' .
  • p51: In Functions 2.1 and 2.2, W=exp(-i... should include the minus sign .
  • p54L-7 to p56: Please note that the notation with f, F, g and G is not consistent with to usual signal name x(t).
  • p66L3: c_1 and c_M should be c[1] and c[M]. Same thing on Line 20 and p67L10 for c[m]
  • p87L3: X should be x in the exponent .
  • p94L3-4: 0.8^|l| should be 0.8^|k| .
  • p95L6: Missing right parenthesis at the end .
  • p98L14: covfun_estimate should be sig2covfun .
  • p101Eq(3.28): Y_N should be Y_T .
  • p109Eq(3.41): \Phi_u should be \Phi_{yy} .
  • p110L-1: Mix of notation, X_N and Y_N should be X_T^(N) and Y_T^(N) .
  • p111Eq(3.43): yu should be ys twice.
  • p113L4-5: yy should be xx three times .
  • p120 Eq(4.2): h(k) should be h[k] and h(0) should be h[0].
  • p172L12: length N should be M for h.
  • p173L15: freq(m) should be ft(m) .
  • p193 Eq(5.43b): H should be C .
  • p193 Eq(5.46): The definition of H is not needed .
  • p194 Eq(5.48): H should be C .
  • p196L-8: w(t)x(t) should be w(t)x^T(t) .
  • p208L14/15: \bar{y} should be y on three places .
  • p215 Item4: The function lss2covfun exists today, with this name. To use a ltf object, first convert to lss
  • p219 Eq(6.6): r^t is 1 when r=1 and can be omitted .
  • p222 Eq(6.22c): Equality sign = missing .
  • p249 Ex 6.10: Ref to 6.10 should be to 6.9 .
  • p259 Item 1: Convention of \Phi mixed up, so every \Phi should be transposed.
  • p276 Eq (7.32b): v(t) should be n(t).
  • p276 Eq (7.33a): v(t) should be w(t).
  • p278 Eq (7.36b): v(t) should be n(t).
  • p278 Eq (7.41-7.42): ee should be nn.
  • p299 Eq (8.7a): w(t)A\hat{x} should be w(t)) = A\hat{x}, so missing parenthesis and equality sign .
  • p302-304: Here K is used for two slightly different quantities. You may change K in (8.20) and onwards to L with the convention that K(t)=AL(t)
  • p309L11: H_{KF} should be H_{e\hat{s}}.
  • p314 Eq (8.38a): v(t) should be w(t). Also, B was not used before, but does not hurt.
  • p315 Alg 8.4 step 1: Second P should not have a bar.
  • p316-317d Eq (8.46): A has a time index here, which was not intended, but does not hurt, the RTS smoother works for time-varying systems.
  • p334 Eq (9.12c): There should be a delay -n_k in all entries with u. Also, on the next line it should be ARX(n_a,n_b,n_k).
  • p340 Eq (9.27c): \varphi should be without transpose.
  • p350 Eq (9.45b): \varepsilon has a missing argument (t) .
  • p361L11: superindex RLS should be WLS .
  • p362: The objects rwarx and jarx are not implemented yet .

Informationsansvarig: Fredrik gustafsson
Senast uppdaterad: 2022-03-18