Least Mean Square (LMS) Filters

void riscv_lms_f32(const riscv_lms_instance_f32 *S, const float32_t *pSrc, float32_t *pRef, float32_t *pOut, float32_t *pErr, uint32_t blockSize)

void riscv_lms_init_f32(riscv_lms_instance_f32 *S, uint16_t numTaps, float32_t *pCoeffs, float32_t *pState, float32_t mu, uint32_t blockSize)

void riscv_lms_init_q15(riscv_lms_instance_q15 *S, uint16_t numTaps, q15_t *pCoeffs, q15_t *pState, q15_t mu, uint32_t blockSize, uint32_t postShift)

void riscv_lms_init_q31(riscv_lms_instance_q31 *S, uint16_t numTaps, q31_t *pCoeffs, q31_t *pState, q31_t mu, uint32_t blockSize, uint32_t postShift)

void riscv_lms_q15(const riscv_lms_instance_q15 *S, const q15_t *pSrc, q15_t *pRef, q15_t *pOut, q15_t *pErr, uint32_t blockSize)

void riscv_lms_q31(const riscv_lms_instance_q31 *S, const q31_t *pSrc, q31_t *pRef, q31_t *pOut, q31_t *pErr, uint32_t blockSize)
 group LMS
LMS filters are a class of adaptive filters that are able to “learn” an unknown transfer functions. LMS filters use a gradient descent method in which the filter coefficients are updated based on the instantaneous error signal. Adaptive filters are often used in communication systems, equalizers, and noise removal. The NMSIS DSP Library contains LMS filter functions that operate on Q15, Q31, and floatingpoint data types. The library also contains normalized LMS filters in which the filter coefficient adaptation is indepedent of the level of the input signal.
An LMS filter consists of two components as shown below. The first component is a standard transversal or FIR filter. The second component is a coefficient update mechanism. The LMS filter has two input signals. The “input” feeds the FIR filter while the “reference input” corresponds to the desired output of the FIR filter. That is, the FIR filter coefficients are updated so that the output of the FIR filter matches the reference input. The filter coefficient update mechanism is based on the difference between the FIR filter output and the reference input. This “error signal” tends towards zero as the filter adapts. The LMS processing functions accept the input and reference input signals and generate the filter output and error signal.
The functions operate on blocks of data and each call to the function processes
blockSize
samples through the filter.pSrc
points to input signal,pRef
points to reference signal,pOut
points to output signal andpErr
points to error signal. All arrays containblockSize
values.The functions operate on a blockbyblock basis. Internally, the filter coefficients
b[n]
are updated on a samplebysample basis. The convergence of the LMS filter is slower compared to the normalized LMS algorithm. Algorithm
The output signal
y[n]
is computed by a standard FIR filter:The error signal equals the difference between the reference signal
d[n]
and the filter output:After each sample of the error signal is computed, the filter coefficients
b[k]
are updated on a samplebysample basis: wheremu
is the step size and controls the rate of coefficient convergence.In the APIs,
pCoeffs
points to a coefficient array of sizenumTaps
. Coefficients are stored in time reversed order.pState
points to a state array of sizenumTaps + blockSize  1
. Samples in the state buffer are stored in the order:Note that the length of the state buffer exceeds the length of the coefficient array by
blockSize1
samples. The increased state buffer length allows circular addressing, which is traditionally used in FIR filters, to be avoided and yields a significant speed improvement. The state variables are updated after each block of data is processed. Instance Structure
The coefficients and state variables for a filter are stored together in an instance data structure. A separate instance structure must be defined for each filter and coefficient and state arrays cannot be shared among instances. There are separate instance structure declarations for each of the 3 supported data types.
 Initialization Functions
There is also an associated initialization function for each data type. The initialization function performs the following operations:
Sets the values of the internal structure fields.
Zeros out the values in the state buffer. To do this manually without calling the init function, assign the follow subfields of the instance structure: numTaps, pCoeffs, mu, postShift (not for f32), pState. Also set all of the values in pState to zero.
Use of the initialization function is optional. However, if the initialization function is used, then the instance structure cannot be placed into a const data section. To place an instance structure into a const data section, the instance structure must be manually initialized. Set the values in the state buffer to zeros before static initialization. The code below statically initializes each of the 3 different data type filter instance structures where
numTaps
is the number of filter coefficients in the filter;pState
is the address of the state buffer;pCoeffs
is the address of the coefficient buffer;mu
is the step size parameter; andpostShift
is the shift applied to coefficients. FixedPoint Behavior
Care must be taken when using the Q15 and Q31 versions of the LMS filter. The following issues must be considered:
Scaling of coefficients
Overflow and saturation
 Scaling of Coefficients
Filter coefficients are represented as fractional values and coefficients are restricted to lie in the range
[1 +1)
. The fixedpoint functions have an additional scaling parameterpostShift
. At the output of the filter’s accumulator is a shift register which shifts the result bypostShift
bits. This essentially scales the filter coefficients by2^postShift
and allows the filter coefficients to exceed the range[+1 1)
. The value ofpostShift
is set by the user based on the expected gain through the system being modeled. Overflow and Saturation
Overflow and saturation behavior of the fixedpoint Q15 and Q31 versions are described separately as part of the function specific documentation below.
Functions

void riscv_lms_f32(const riscv_lms_instance_f32 *S, const float32_t *pSrc, float32_t *pRef, float32_t *pOut, float32_t *pErr, uint32_t blockSize)
Processing function for floatingpoint LMS filter.
 Parameters
S – [in] points to an instance of the floatingpoint LMS filter structure
pSrc – [in] points to the block of input data
pRef – [in] points to the block of reference data
pOut – [out] points to the block of output data
pErr – [out] points to the block of error data
blockSize – [in] number of samples to process
 Returns
none

void riscv_lms_init_f32(riscv_lms_instance_f32 *S, uint16_t numTaps, float32_t *pCoeffs, float32_t *pState, float32_t mu, uint32_t blockSize)
Initialization function for floatingpoint LMS filter.
 Details
pCoeffs
points to the array of filter coefficients stored in time reversed order: The initial filter coefficients serve as a starting point for the adaptive filter.pState
points to an array of lengthnumTaps+blockSize1
samples, whereblockSize
is the number of input samples processed by each call toriscv_lms_f32()
.
 Parameters
S – [in] points to an instance of the floatingpoint LMS filter structure
numTaps – [in] number of filter coefficients
pCoeffs – [in] points to coefficient buffer
pState – [in] points to state buffer
mu – [in] step size that controls filter coefficient updates
blockSize – [in] number of samples to process
 Returns
none

void riscv_lms_init_q15(riscv_lms_instance_q15 *S, uint16_t numTaps, q15_t *pCoeffs, q15_t *pState, q15_t mu, uint32_t blockSize, uint32_t postShift)
Initialization function for the Q15 LMS filter.
 Details
pCoeffs
points to the array of filter coefficients stored in time reversed order: The initial filter coefficients serve as a starting point for the adaptive filter.pState
points to the array of state variables and size of array isnumTaps+blockSize1
samples, whereblockSize
is the number of input samples processed by each call toriscv_lms_q15()
.
 Parameters
S – [in] points to an instance of the Q15 LMS filter structure.
numTaps – [in] number of filter coefficients.
pCoeffs – [in] points to coefficient buffer.
pState – [in] points to state buffer.
mu – [in] step size that controls filter coefficient updates.
blockSize – [in] number of samples to process.
postShift – [in] bit shift applied to coefficients.
 Returns
none

void riscv_lms_init_q31(riscv_lms_instance_q31 *S, uint16_t numTaps, q31_t *pCoeffs, q31_t *pState, q31_t mu, uint32_t blockSize, uint32_t postShift)
Initialization function for Q31 LMS filter.
 Details
pCoeffs
points to the array of filter coefficients stored in time reversed order: The initial filter coefficients serve as a starting point for the adaptive filter.pState
points to an array of lengthnumTaps+blockSize1
samples, whereblockSize
is the number of input samples processed by each call toriscv_lms_q31()
.
 Parameters
S – [in] points to an instance of the Q31 LMS filter structure
numTaps – [in] number of filter coefficients
pCoeffs – [in] points to coefficient buffer
pState – [in] points to state buffer
mu – [in] step size that controls filter coefficient updates
blockSize – [in] number of samples to process
postShift – [in] bit shift applied to coefficients
 Returns
none

void riscv_lms_q15(const riscv_lms_instance_q15 *S, const q15_t *pSrc, q15_t *pRef, q15_t *pOut, q15_t *pErr, uint32_t blockSize)
Processing function for Q15 LMS filter.
 Scaling and Overflow Behavior
The function is implemented using an internal 64bit accumulator. Both coefficients and state variables are represented in 1.15 format and multiplications yield a 2.30 result. The 2.30 intermediate results are accumulated in a 64bit accumulator in 34.30 format. There is no risk of internal overflow with this approach and the full precision of intermediate multiplications is preserved. After all additions have been performed, the accumulator is truncated to 34.15 format by discarding low 15 bits. Lastly, the accumulator is saturated to yield a result in 1.15 format.
In this filter, filter coefficients are updated for each sample and the updation of filter cofficients are saturted.
 Parameters
S – [in] points to an instance of the Q15 LMS filter structure
pSrc – [in] points to the block of input data
pRef – [in] points to the block of reference data
pOut – [out] points to the block of output data
pErr – [out] points to the block of error data
blockSize – [in] number of samples to process
 Returns
none

void riscv_lms_q31(const riscv_lms_instance_q31 *S, const q31_t *pSrc, q31_t *pRef, q31_t *pOut, q31_t *pErr, uint32_t blockSize)
Processing function for Q31 LMS filter.
 Scaling and Overflow Behavior
The function is implemented using an internal 64bit accumulator. The accumulator has a 2.62 format and maintains full precision of the intermediate multiplication results but provides only a single guard bit. Thus, if the accumulator result overflows it wraps around rather than clips. In order to avoid overflows completely the input signal must be scaled down by log2(numTaps) bits. The reference signal should not be scaled down. After all multiplyaccumulates are performed, the 2.62 accumulator is shifted and saturated to 1.31 format to yield the final result. The output signal and error signal are in 1.31 format.
In this filter, filter coefficients are updated for each sample and the updation of filter cofficients are saturted.
 Parameters
S – [in] points to an instance of the Q31 LMS filter structure.
pSrc – [in] points to the block of input data.
pRef – [in] points to the block of reference data.
pOut – [out] points to the block of output data.
pErr – [out] points to the block of error data.
blockSize – [in] number of samples to process.
 Returns
none