Data Type Agent

Recommends fixed-point data types for Fixed-Point Designer blocks

Since R2025a

Description

The Data Type Agent app recommends fixed-point data types for Simulink^® blocks in the Fixed-Point Designer™ block libraries. The Data Type Agent app supports these blocks:

• Real Partial-Systolic Matrix Solve Using QR Decomposition

• Complex Partial-Systolic Matrix Solve Using QR Decomposition

Open the Data Type Agent

• Simulink Toolstrip: Click a supported block in the model. On the Block tab, under Data Type, click the app icon.

Examples

• Determine Fixed-Point Types for Real Least-Squares Matrix Solve AX=B
• Determine Fixed-Point Types for Real Least-Squares Matrix Solve with Tikhonov Regularization
• Determine Fixed-Point Types for Complex Least-Squares Matrix Solve AX=B
• Determine Fixed-Point Types for Complex Least-Squares Matrix Solve with Tikhonov Regularization

Parameters

Block Parameters

Number of rows in matrices A and B — Number of rows in matrices A and B

m (default) | positive integer-valued scalar

Number of rows in matrices A and B, specified as a positive integer-valued scalar. By default, this parameter is set to the value of the block parameter m.

Number of columns in matrix A — Number of columns in matrix A

n (default) | positive integer-valued scalar

Number of columns in matrix A, specified as a positive integer-valued scalar. By default, this parameter is set to the value of the block parameter n.

Number of columns in matrix B — Number of columns in matrix B

p (default) | positive integer-valued scalar

Number of columns in matrix B, specified as a positive integer-valued scalar. By default, this parameter is set to the value of the block parameter p.

Regularization Parameter — Regularization parameter

regularizationParameter (default) | real-valued nonnegative scalar

Regularization parameter, specified as a real-valued nonnegative scalar. By default, this parameter is set to the value of the block parameter regularizationParameter.

Precision Bits — Number of bits of precision required

precision specified by OutputType (default) | positive integer-valued scalar

Number of bits of precision required for the matrix solve, specified as a positive integer-valued scalar. Set this value according to system requirements. By default, this parameter is set to the precision specified by the block parameter OutputType.

Additional Parameters

Max of absolute value of A — Upper bound on maximum magnitude element of A

scalar

Upper bound on the maximum magnitude element of matrix A, specified as a scalar.

Max of absolute value of B — Upper bound on maximum magnitude element of B

scalar

Upper bound on the maximum magnitude element of matrix B, specified as a scalar.

Noise Standard Deviation — Estimate of standard deviation of quantization noise

valued determined by fixed.realQuantizationNoiseStandardDeviation or fixed.complexQuantizationNoiseStandardDeviation (default) | scalar

Estimate of the standard deviation of the noise resulting from quantizing the elements of a real signal, specified as a scalar. The quantization noise standard deviation is a function of the required number of bits of precision. A well-designed system has the quantization noise standard deviation level lower than the thermal noise standard deviation. By default, the Estimate Standard Deviation from Quantization Noise parameter is set to on and the app automatically determines the noise standard deviation by using the fixed.realQuantizationNoiseStandardDeviation or fixed.complexQuantizationNoiseStandardDeviation function.

Estimate Standard Deviation from Quantization Noise — Whether to estimate the noise standard deviation from quantization noise

on (default) | off

Whether to estimate the Noise Standard Deviation parameter from quantization noise. By default, this parameter is set to on and the app determines the noise standard deviation by using the fixed.realQuantizationNoiseStandardDeviation or fixed.complexQuantizationNoiseStandardDeviation function.

Overflow Probability — Probability of overflow

scalar

Probability of overflow, specified as a scalar. By default, the app determines fixed-point data types such that the probability of overflow is low.

Max Word Length — Maximum allowed word length

128 (default) | positive integer-valued scalar

Maximum allowed word length, specified as a positive integer-valued scalar.

Limitations

• When the specified Overflow Probability is small, the algorithms used by the tool to estimate the probability of overflow perform well and the estimated upper bound of overflow probability is close to the true overflow probability. When the Overflow Probability is large, the estimated upper bound of overflow probability diverges and the algorithms used by the tool may produce less accurate results.

Algorithms

The Data Type Agent provides an interactive visualization which is equivalent to manually determining fixed-point data types using the fixed.realQRMatrixSolveFixedpointTypes or fixed.complexQRMatrixSolveFixedpointTypes functions.

For a full description of the algorithms used by the Data Type Agent, see:

• Algorithms to Determine Fixed-Point Types for Real Least-Squares Matrix Solve AX=B

• Algorithms to Determine Fixed-Point Types for Complex Least-Squares Matrix Solve AX=B

Version History

Introduced in R2025a