The Characteristic Equation and Stability in Control Systems

The Routh-Hurwitz criterion is a cornerstone of classical control theory, providing a powerful algebraic method to determine the stability of a linear time-invariant (LTI) system without explicitly computing its poles. For systems described by a fourth-order polynomial characteristic equation, this tool becomes invaluable because solving quartic equations analytically can be tedious and time-consuming. Stability, in this context, means that the system's response to bounded inputs remains bounded over time—a condition that corresponds to all poles of the transfer function having negative real parts. The Routh-Hurwitz criterion allows engineers to verify this condition by inspecting the signs of a specially constructed array known as the Routh array.

Form of the Fourth-Order Polynomial

The general form of a fourth-order characteristic equation is:

a₄s⁴ + a₃s³ + a₂s² + a₁s + a₀ = 0

Here, the coefficients a₄ through a₀ are real constants, and a₄ is typically positive (if negative, the entire equation can be multiplied by -1 without affecting the root locations). The order of the polynomial determines the number of roots (poles), and for a strictly stable system, all roots must lie in the left half of the complex s-plane (Re(s) < 0). For a fourth-order polynomial, this means five coefficients must satisfy certain inequalities—conditions that the Routh-Hurwitz criterion organizes into a simple tabular test.

Necessary Conditions for Stability

Before constructing the Routh array, it is wise to check the necessary (but not sufficient) conditions: all coefficients must be positive and non-zero. For the polynomial given above, this means:

  • a₄ > 0
  • a₃ > 0
  • a₂ > 0
  • a₁ > 0
  • a₀ > 0

If any coefficient is zero or negative, the system is either unstable or marginally stable (roots on the imaginary axis). However, these conditions alone are not enough; a polynomial with all positive coefficients can still have right-half-plane roots. The full Routh-Hurwitz procedure provides the sufficient conditions by constructing the array and examining sign changes in the first column.

Step-by-Step Construction of the Routh Array

The Routh array for a fourth-order polynomial is a systematic arrangement of coefficients that allows the stability test to be performed through simple arithmetic. The procedure is purely deterministic and can be extended to polynomials of any order, but the fourth-order case is common enough to warrant a detailed walkthrough.

Step 1: Arrange the First Two Rows

The first two rows of the array are formed directly from the coefficients of the characteristic equation. The pattern is as follows:

  • Row s⁴ (first row): Contains coefficients of even powers: a₄, a₂, a₀ (in that order). If the polynomial lacks a term, insert a zero.
  • Row s³ (second row): Contains coefficients of odd powers: a₃, a₁, 0 (the last entry is zero because there is no s-1 term).

Thus for our general fourth-order polynomial, the array begins as:

s⁴:  a₄   a₂   a₀
s³:  a₃   a₁   0

Note that the number of columns is determined by the highest order; for a fourth-order system, we need three columns to accommodate all coefficients. The first two rows are always populated directly from the polynomial.

Step 2: Compute the Remaining Rows

Subsequent rows are labeled s², s¹, and s⁰ (the constant term row). The elements of each row are computed using the determinant formula:

bₖₗ = - (1 / row_above_leftmost) × det( [ previous_row_left previous_row_next ; two_rows_above_left two_rows_above_next] )

Where "row_above_leftmost" refers to the element in the same column of the row immediately above the one being computed. In practice, for the s² row, we take the first two elements from the s³ row (first column as pivot) and the first two elements from the s⁴ row to form a 2x2 determinant.

Detailed Formulas for a Fourth-Order Polynomial

Let’s define the first two rows more precisely:

  • Row s⁴: c₁ = a₄, c₂ = a₂, c₃ = a₀
  • Row s³: d₁ = a₃, d₂ = a₁, d₃ = 0

Then the s² row (call it e₁, e₂, e₃) is computed as:

  • e₁ = (d₁ × c₂ - c₁ × d₂) / d₁
  • e₂ = (d₁ × c₃ - c₁ × d₃) / d₁ (note d₃ = 0, so this simplifies to c₃)
  • e₃ = 0 (always, since the determinant leading to the third column will involve zeros)

After computing the s² row, we compute the s¹ row (call it f₁, f₂) using the two rows above it: s³ and s². The pivot element is the first element of the s² row (e₁).

  • f₁ = (e₁ × d₂ - d₁ × e₂) / e₁
  • f₂ = (e₁ × d₃ - d₁ × e₃) / e₁ = 0 (since d₃=0, e₃=0)

Finally, row s⁰ (call it g₁) comes from the s¹ and s² rows:

  • g₁ = (f₁ × e₂ - e₁ × f₂) / f₁ (f₂ = 0, so this equals e₂)

Thus the completed Routh array for a typical fourth-order system has five rows: s⁴, s³, s², s¹, s⁰. All cells in each row are computed sequentially; if any pivot becomes zero, special cases (discussed later) arise.

Worked Example

Consider the polynomial: 2s⁴ + 3s³ + 4s² + 5s + 6 = 0.

The first two rows are:

  • s⁴: 2, 4, 6
  • s³: 3, 5, 0

Now compute the s² row:

  • e₁ = (3×4 - 2×5) / 3 = (12 - 10) / 3 = 2/3 ≈ 0.667
  • e₂ = (3×6 - 2×0) / 3 = 18/3 = 6
  • e₃ = 0

s¹ row:

  • f₁ = (e₁×5 - 3×e₂) / e₁ = ((2/3)×5 - 3×6) / (2/3) = (10/3 - 18) / (2/3) = ( (10 - 54)/3 ) / (2/3) = (-44/3) × (3/2) = -44/2 = -22
  • f₂ = 0

s⁰ row:

  • g₁ = (f₁×e₂ - e₁×f₂) / f₁ = (-22×6 - (2/3)×0) / (-22) = (-132) / (-22) = 6

The completed array is:

s⁴:  2      4      6
s³:  3      5      0
s²:   2/3    6      0
s¹:  -22     0
s⁰:  6

Now we examine the first column: [2, 3, 2/3, -22, 6]. There is a sign change from 2/3 (positive) to -22 (negative), and then back to 6 (positive). This indicates two sign changes, meaning the polynomial has two roots with positive real parts (unstable). Indeed, solving the quartic would confirm instability.

Interpreting the Routh Array for Stability

The First-Column Sign Rule

The fundamental stability criterion is: a system is stable if and only if all elements in the first column of the Routh array are positive (and non-zero). If any element is negative, the number of sign changes equals the number of roots in the right half-plane. If an element is zero, the system is marginally stable or the array requires special handling (see below). For the fourth-order case, stability requires all five entries in the first column (rows s⁴ through s⁰) to be > 0.

Special Cases: Zero in the First Column or a Row of Zeros

Two common special circumstances can occur when constructing the Routh array for a fourth-order system:

  1. Zero in the first column, but not an entire row of zeros: This can happen when a pivot becomes zero, making the next determinant undefined. The standard technique is to replace the zero with a small positive number ε (epsilon), complete the array using ε, and then examine the signs as ε → 0⁺. If the sign changes after replacing with ε, the root count is determined by the limiting behavior. Alternatively, the polynomial can be multiplied by (s + a) where a > 0 to shift the roots slightly, but the ε method is more straightforward for manual calculation.
  2. Entire row of zeros: This occurs when the polynomial has a pair of roots symmetric about the origin (e.g., pure imaginary roots or symmetric real roots). In that case, an auxiliary polynomial is formed from the row immediately above the all-zero row. The derivative of that auxiliary polynomial provides coefficients for the zero row, and the array can be completed. The presence of a row of zeros indicates marginal stability or instability, depending on the roots of the auxiliary polynomial. For a fourth-order system, a row of zeros at the s² row would imply two pairs of symmetric roots (e.g., two imaginary conjugate pairs).

Both cases require careful handling, and the Routh-Hurwitz criterion remains applicable provided the special procedures are correctly applied. Many textbooks and online resources provide detailed examples of these special cases, such as the Wikipedia article on Routh-Hurwitz criterion and Javatpoint's control systems tutorial.

Practical Applications and Limitations

Advantages in Control System Design

The Routh-Hurwitz criterion is widely used because it avoids the computational burden of factoring high-order polynomials. For a fourth-order system, the array can be constructed by hand in minutes, providing immediate insight into stability without resorting to numerical root-finding algorithms. The method also allows engineers to determine the effect of varying a parameter on stability by treating that parameter symbolically in the array. For example, in tuning a PID controller for a fourth-order plant, the characteristic equation coefficients become functions of the controller gains, and the Routh array yields inequalities that guarantee stability—often called the Routh-Hurwitz stability margins. This symbolic approach is invaluable during the design phase, as explained in Control Systems Academy's resources.

Limitations and Complementary Methods

Despite its elegance, the Routh-Hurwitz criterion has limitations. It provides only a yes/no answer on stability (or the number of unstable poles) but does not give the actual pole locations or transient response characteristics. For design purposes, engineers often combine the Routh test with root locus or Bode plots. Additionally, the criterion assumes rational transfer functions with real coefficients; it does not handle dead-time or nonlinearities. For high-order systems (e.g., tenth-order or higher), hand calculation becomes cumbersome, though software tools like MATLAB can automate the array construction using the routh command or the symbolic toolbox.

Another practical limitation: numerical precision matters when coefficients are close to zero; small rounding errors can change the sign of a first-column element, leading to an incorrect stability conclusion. Therefore, engineers should verify borderline cases using alternative methods such as the Nyquist criterion or eigenvalue computation. For a thorough understanding of linear system stability, ScienceDirect's topic page on the Routh-Hurwitz criterion provides additional theoretical background and references.

Conclusion

The step-by-step Routh-Hurwitz analysis for a fourth-order polynomial system is a systematic, reliable method for assessing stability without solving the characteristic equation. By constructing the array from the coefficients and checking the signs of the first column, engineers can quickly determine whether all poles lie in the left half-plane. Special cases—zero in the first column or rows of zeros—are handled with well-known extensions to the basic procedure. While the method has limitations and does not replace numerical simulation or frequency-domain analysis, it remains an essential skill for control engineers and a practical tool for designing stable feedback systems. With the provided formulas and worked example, you can confidently apply the Routh-Hurwitz criterion to any fourth-order system.