123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442 |
- /* ----------------------------------------------------------------------
- * Copyright (C) 2010-2013 ARM Limited. All rights reserved.
- *
- * $Date: 17. January 2013
- * $Revision: V1.4.1
- *
- * Project: CMSIS DSP Library
- * Title: arm_lms_f32.c
- *
- * Description: Processing function for the floating-point LMS filter.
- *
- * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * - Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- * - Redistributions in binary form must reproduce the above copyright
- * notice, this list of conditions and the following disclaimer in
- * the documentation and/or other materials provided with the
- * distribution.
- * - Neither the name of ARM LIMITED nor the names of its contributors
- * may be used to endorse or promote products derived from this
- * software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
- * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
- * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
- * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
- * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
- * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
- * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
- * POSSIBILITY OF SUCH DAMAGE.
- * -------------------------------------------------------------------- */
-
- #include "arm_math.h"
-
- /**
- * @ingroup groupFilters
- */
-
- /**
- * @defgroup LMS Least Mean Square (LMS) Filters
- *
- * 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 CMSIS DSP Library contains LMS filter functions that operate on Q15, Q31, and floating-point 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.
- * \image html LMS.gif "Internal structure of the Least Mean Square filter"
- *
- * The functions operate on blocks of data and each call to the function processes
- * <code>blockSize</code> samples through the filter.
- * <code>pSrc</code> points to input signal, <code>pRef</code> points to reference signal,
- * <code>pOut</code> points to output signal and <code>pErr</code> points to error signal.
- * All arrays contain <code>blockSize</code> values.
- *
- * The functions operate on a block-by-block basis.
- * Internally, the filter coefficients <code>b[n]</code> are updated on a sample-by-sample basis.
- * The convergence of the LMS filter is slower compared to the normalized LMS algorithm.
- *
- * \par Algorithm:
- * The output signal <code>y[n]</code> is computed by a standard FIR filter:
- * <pre>
- * y[n] = b[0] * x[n] + b[1] * x[n-1] + b[2] * x[n-2] + ...+ b[numTaps-1] * x[n-numTaps+1]
- * </pre>
- *
- * \par
- * The error signal equals the difference between the reference signal <code>d[n]</code> and the filter output:
- * <pre>
- * e[n] = d[n] - y[n].
- * </pre>
- *
- * \par
- * After each sample of the error signal is computed, the filter coefficients <code>b[k]</code> are updated on a sample-by-sample basis:
- * <pre>
- * b[k] = b[k] + e[n] * mu * x[n-k], for k=0, 1, ..., numTaps-1
- * </pre>
- * where <code>mu</code> is the step size and controls the rate of coefficient convergence.
- *\par
- * In the APIs, <code>pCoeffs</code> points to a coefficient array of size <code>numTaps</code>.
- * Coefficients are stored in time reversed order.
- * \par
- * <pre>
- * {b[numTaps-1], b[numTaps-2], b[N-2], ..., b[1], b[0]}
- * </pre>
- * \par
- * <code>pState</code> points to a state array of size <code>numTaps + blockSize - 1</code>.
- * Samples in the state buffer are stored in the order:
- * \par
- * <pre>
- * {x[n-numTaps+1], x[n-numTaps], x[n-numTaps-1], x[n-numTaps-2]....x[0], x[1], ..., x[blockSize-1]}
- * </pre>
- * \par
- * Note that the length of the state buffer exceeds the length of the coefficient array by <code>blockSize-1</code> 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.
- * \par 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.
- *
- * \par 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.
- *
- * \par
- * 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
- * <pre>
- * arm_lms_instance_f32 S = {numTaps, pState, pCoeffs, mu};
- * arm_lms_instance_q31 S = {numTaps, pState, pCoeffs, mu, postShift};
- * arm_lms_instance_q15 S = {numTaps, pState, pCoeffs, mu, postShift};
- * </pre>
- * where <code>numTaps</code> is the number of filter coefficients in the filter; <code>pState</code> is the address of the state buffer;
- * <code>pCoeffs</code> is the address of the coefficient buffer; <code>mu</code> is the step size parameter; and <code>postShift</code> is the shift applied to coefficients.
- *
- * \par Fixed-Point 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
- *
- * \par Scaling of Coefficients:
- * Filter coefficients are represented as fractional values and
- * coefficients are restricted to lie in the range <code>[-1 +1)</code>.
- * The fixed-point functions have an additional scaling parameter <code>postShift</code>.
- * At the output of the filter's accumulator is a shift register which shifts the result by <code>postShift</code> bits.
- * This essentially scales the filter coefficients by <code>2^postShift</code> and
- * allows the filter coefficients to exceed the range <code>[+1 -1)</code>.
- * The value of <code>postShift</code> is set by the user based on the expected gain through the system being modeled.
- *
- * \par Overflow and Saturation:
- * Overflow and saturation behavior of the fixed-point Q15 and Q31 versions are
- * described separately as part of the function specific documentation below.
- */
-
- /**
- * @addtogroup LMS
- * @{
- */
-
- /**
- * @details
- * This function operates on floating-point data types.
- *
- * @brief Processing function for floating-point LMS filter.
- * @param[in] *S points to an instance of the floating-point LMS filter structure.
- * @param[in] *pSrc points to the block of input data.
- * @param[in] *pRef points to the block of reference data.
- * @param[out] *pOut points to the block of output data.
- * @param[out] *pErr points to the block of error data.
- * @param[in] blockSize number of samples to process.
- * @return none.
- */
-
- void arm_lms_f32(
- const arm_lms_instance_f32 * S,
- float32_t * pSrc,
- float32_t * pRef,
- float32_t * pOut,
- float32_t * pErr,
- uint32_t blockSize)
- {
- float32_t *pState = S->pState; /* State pointer */
- float32_t *pCoeffs = S->pCoeffs; /* Coefficient pointer */
- float32_t *pStateCurnt; /* Points to the current sample of the state */
- float32_t *px, *pb; /* Temporary pointers for state and coefficient buffers */
- float32_t mu = S->mu; /* Adaptive factor */
- uint32_t numTaps = S->numTaps; /* Number of filter coefficients in the filter */
- uint32_t tapCnt, blkCnt; /* Loop counters */
- float32_t sum, e, d; /* accumulator, error, reference data sample */
- float32_t w = 0.0f; /* weight factor */
-
- e = 0.0f;
- d = 0.0f;
-
- /* S->pState points to state array which contains previous frame (numTaps - 1) samples */
- /* pStateCurnt points to the location where the new input data should be written */
- pStateCurnt = &(S->pState[(numTaps - 1u)]);
-
- blkCnt = blockSize;
-
-
- #ifndef ARM_MATH_CM0_FAMILY
-
- /* Run the below code for Cortex-M4 and Cortex-M3 */
-
- while(blkCnt > 0u)
- {
- /* Copy the new input sample into the state buffer */
- *pStateCurnt++ = *pSrc++;
-
- /* Initialize pState pointer */
- px = pState;
-
- /* Initialize coeff pointer */
- pb = (pCoeffs);
-
- /* Set the accumulator to zero */
- sum = 0.0f;
-
- /* Loop unrolling. Process 4 taps at a time. */
- tapCnt = numTaps >> 2;
-
- while(tapCnt > 0u)
- {
- /* Perform the multiply-accumulate */
- sum += (*px++) * (*pb++);
- sum += (*px++) * (*pb++);
- sum += (*px++) * (*pb++);
- sum += (*px++) * (*pb++);
-
- /* Decrement the loop counter */
- tapCnt--;
- }
-
- /* If the filter length is not a multiple of 4, compute the remaining filter taps */
- tapCnt = numTaps % 0x4u;
-
- while(tapCnt > 0u)
- {
- /* Perform the multiply-accumulate */
- sum += (*px++) * (*pb++);
-
- /* Decrement the loop counter */
- tapCnt--;
- }
-
- /* The result in the accumulator, store in the destination buffer. */
- *pOut++ = sum;
-
- /* Compute and store error */
- d = (float32_t) (*pRef++);
- e = d - sum;
- *pErr++ = e;
-
- /* Calculation of Weighting factor for the updating filter coefficients */
- w = e * mu;
-
- /* Initialize pState pointer */
- px = pState;
-
- /* Initialize coeff pointer */
- pb = (pCoeffs);
-
- /* Loop unrolling. Process 4 taps at a time. */
- tapCnt = numTaps >> 2;
-
- /* Update filter coefficients */
- while(tapCnt > 0u)
- {
- /* Perform the multiply-accumulate */
- *pb = *pb + (w * (*px++));
- pb++;
-
- *pb = *pb + (w * (*px++));
- pb++;
-
- *pb = *pb + (w * (*px++));
- pb++;
-
- *pb = *pb + (w * (*px++));
- pb++;
-
- /* Decrement the loop counter */
- tapCnt--;
- }
-
- /* If the filter length is not a multiple of 4, compute the remaining filter taps */
- tapCnt = numTaps % 0x4u;
-
- while(tapCnt > 0u)
- {
- /* Perform the multiply-accumulate */
- *pb = *pb + (w * (*px++));
- pb++;
-
- /* Decrement the loop counter */
- tapCnt--;
- }
-
- /* Advance state pointer by 1 for the next sample */
- pState = pState + 1;
-
- /* Decrement the loop counter */
- blkCnt--;
- }
-
-
- /* Processing is complete. Now copy the last numTaps - 1 samples to the
- satrt of the state buffer. This prepares the state buffer for the
- next function call. */
-
- /* Points to the start of the pState buffer */
- pStateCurnt = S->pState;
-
- /* Loop unrolling for (numTaps - 1u) samples copy */
- tapCnt = (numTaps - 1u) >> 2u;
-
- /* copy data */
- while(tapCnt > 0u)
- {
- *pStateCurnt++ = *pState++;
- *pStateCurnt++ = *pState++;
- *pStateCurnt++ = *pState++;
- *pStateCurnt++ = *pState++;
-
- /* Decrement the loop counter */
- tapCnt--;
- }
-
- /* Calculate remaining number of copies */
- tapCnt = (numTaps - 1u) % 0x4u;
-
- /* Copy the remaining q31_t data */
- while(tapCnt > 0u)
- {
- *pStateCurnt++ = *pState++;
-
- /* Decrement the loop counter */
- tapCnt--;
- }
-
- #else
-
- /* Run the below code for Cortex-M0 */
-
- while(blkCnt > 0u)
- {
- /* Copy the new input sample into the state buffer */
- *pStateCurnt++ = *pSrc++;
-
- /* Initialize pState pointer */
- px = pState;
-
- /* Initialize pCoeffs pointer */
- pb = pCoeffs;
-
- /* Set the accumulator to zero */
- sum = 0.0f;
-
- /* Loop over numTaps number of values */
- tapCnt = numTaps;
-
- while(tapCnt > 0u)
- {
- /* Perform the multiply-accumulate */
- sum += (*px++) * (*pb++);
-
- /* Decrement the loop counter */
- tapCnt--;
- }
-
- /* The result is stored in the destination buffer. */
- *pOut++ = sum;
-
- /* Compute and store error */
- d = (float32_t) (*pRef++);
- e = d - sum;
- *pErr++ = e;
-
- /* Weighting factor for the LMS version */
- w = e * mu;
-
- /* Initialize pState pointer */
- px = pState;
-
- /* Initialize pCoeffs pointer */
- pb = pCoeffs;
-
- /* Loop over numTaps number of values */
- tapCnt = numTaps;
-
- while(tapCnt > 0u)
- {
- /* Perform the multiply-accumulate */
- *pb = *pb + (w * (*px++));
- pb++;
-
- /* Decrement the loop counter */
- tapCnt--;
- }
-
- /* Advance state pointer by 1 for the next sample */
- pState = pState + 1;
-
- /* Decrement the loop counter */
- blkCnt--;
- }
-
-
- /* Processing is complete. Now copy the last numTaps - 1 samples to the
- * start of the state buffer. This prepares the state buffer for the
- * next function call. */
-
- /* Points to the start of the pState buffer */
- pStateCurnt = S->pState;
-
- /* Copy (numTaps - 1u) samples */
- tapCnt = (numTaps - 1u);
-
- /* Copy the data */
- while(tapCnt > 0u)
- {
- *pStateCurnt++ = *pState++;
-
- /* Decrement the loop counter */
- tapCnt--;
- }
-
- #endif /* #ifndef ARM_MATH_CM0_FAMILY */
-
- }
-
- /**
- * @} end of LMS group
- */
|