osmo-bts/tests/power/ms_power_loop_test.c

434 lines
16 KiB
C
Raw Normal View History

/*
* (C) 2013,2014 by Holger Hans Peter Freyther
* Contributions by sysmocom - s.m.f.c. GmbH <info@sysmocom.de>
*
* All Rights Reserved
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as published by
* the Free Software Foundation; either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <osmocom/core/talloc.h>
#include <osmocom/core/application.h>
#include <osmo-bts/bts.h>
#include <osmo-bts/logging.h>
#include <osmo-bts/l1sap.h>
#include <osmo-bts/power_control.h>
#include <stdio.h>
#define PWR_TEST_RXLEV_TARGET_DBM -75
#define PWR_TEST_RXLEV_TARGET \
dbm2rxlev(PWR_TEST_RXLEV_TARGET_DBM)
static struct gsm_bts *g_bts = NULL;
static struct gsm_bts_trx *g_trx = NULL;
static void init_test(const char *name)
{
if (g_trx != NULL)
talloc_free(g_trx);
if (g_bts != NULL)
talloc_free(g_bts);
g_bts = talloc_zero(tall_bts_ctx, struct gsm_bts);
OSMO_ASSERT(g_bts != NULL);
INIT_LLIST_HEAD(&g_bts->trx_list);
g_trx = gsm_bts_trx_alloc(g_bts);
OSMO_ASSERT(g_trx != NULL);
g_trx->ms_pwr_ctl_soft = true;
g_bts->band = GSM_BAND_1800;
g_bts->c0 = g_trx;
/* Init default MS power control parameters, enable dynamic power control */
struct gsm_power_ctrl_params *params = &g_trx->ts[0].lchan[0].ms_dpc_params;
g_trx->ts[0].lchan[0].ms_power_ctrl.dpc_params = params;
*params = power_ctrl_params_def;
/* Disable loop SACCH block skip by default: */
params->ctrl_interval = 0;
/* Disable RxLev pre-processing and hysteresis by default */
struct gsm_power_ctrl_meas_params *mp = &params->rxlev_meas;
mp->lower_thresh = mp->upper_thresh = PWR_TEST_RXLEV_TARGET;
mp->algo = GSM_PWR_CTRL_MEAS_AVG_ALGO_NONE;
printf("\nStarting test case '%s'\n", name);
}
static void apply_power_test_ext(struct gsm_lchan *lchan, uint8_t ms_pwr, int rxlev, int lqual_cb, int exp_ret, uint8_t exp_current)
{
int ret;
ret = lchan_ms_pwr_ctrl(lchan, ms_pwr, rxlev, lqual_cb);
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
/* Keep the measurement counter updated */
lchan->meas.res_nr++;
printf("lchan_ms_pwr_ctrl(RxLvl=%d dBm) returns %d (expected %d)\n",
rxlev, ret, exp_ret);
printf("\tMS current power %u -> %u (expected %u)\n",
ms_pwr, lchan->ms_power_ctrl.current, exp_current);
}
static inline void apply_power_test(struct gsm_lchan *lchan, int rxlev, int lqual_cb, int exp_ret, uint8_t exp_current)
{
apply_power_test_ext(lchan, lchan->ms_power_ctrl.current, rxlev, lqual_cb, exp_ret, exp_current);
}
static void test_power_loop(void)
{
struct gsm_lchan *lchan;
const struct gsm_power_ctrl_params *params;
int16_t good_lqual;
init_test(__func__);
lchan = &g_trx->ts[0].lchan[0];
params = lchan->ms_power_ctrl.dpc_params;
lchan->type = GSM_LCHAN_SDCCH;
good_lqual = (params->ci_sdcch_meas.lower_thresh + 2) * 10;
lchan->ms_power_ctrl.current = ms_pwr_ctl_lvl(GSM_BAND_1800, 0);
OSMO_ASSERT(lchan->ms_power_ctrl.current == 15);
lchan->ms_power_ctrl.max = ms_pwr_ctl_lvl(GSM_BAND_1800, 26);
OSMO_ASSERT(lchan->ms_power_ctrl.max == 2);
/* Simply clamping */
apply_power_test(lchan, -60, good_lqual, 0, 15);
/*
* Now 15 dB too little and we should power it up. Could be a
* power level of 7 or 8 for 15 dBm. However, since we limit peace at
* which we change values, expect several steps of MS_RAISE_MAX_DB/2 levels:
*/
apply_power_test(lchan, -90, good_lqual, 1, 13);
apply_power_test(lchan, -90, good_lqual, 1, 11);
apply_power_test(lchan, -90, good_lqual, 1, 9);
apply_power_test(lchan, -90, good_lqual, 1, 7);
apply_power_test(lchan, -90, good_lqual, 1, 5);
/* Check good RSSI value keeps it at same power level: */
apply_power_test(lchan, PWR_TEST_RXLEV_TARGET_DBM, good_lqual, 0, 5);
apply_power_test(lchan, -90, good_lqual, 1, 3);
apply_power_test(lchan, -90, good_lqual, 1, 2); /* .max is pwr lvl 2 */
apply_power_test(lchan, -90, good_lqual, 0, 2); /* .max is pwr lvl 2 */
lchan->ms_power_ctrl.max = ms_pwr_ctl_lvl(GSM_BAND_1800, 30);
OSMO_ASSERT(lchan->ms_power_ctrl.max == 0);
apply_power_test(lchan, -90, good_lqual, 1, 0); /* .max is pwr lvl 0 */
apply_power_test(lchan, -90, good_lqual, 0, 0); /* .max is pwr lvl 0 */
lchan->ms_power_ctrl.max = ms_pwr_ctl_lvl(GSM_BAND_1800, 36);
OSMO_ASSERT(lchan->ms_power_ctrl.max == 29);
apply_power_test(lchan, -90, good_lqual, 1, 30);
apply_power_test(lchan, -90, good_lqual, 1, 29);
apply_power_test(lchan, -90, good_lqual, 0, 29);
/* Check good RSSI value keeps it at same power level: */
apply_power_test(lchan, PWR_TEST_RXLEV_TARGET_DBM, good_lqual, 0, 29);
/* Now go down, steps are double size in this direction: */
apply_power_test(lchan, -45, good_lqual, 1, 1);
apply_power_test(lchan, -45, good_lqual, 1, 5);
apply_power_test(lchan, -45, good_lqual, 1, 9);
/* Go down only one level down and up: */
apply_power_test(lchan, PWR_TEST_RXLEV_TARGET_DBM + 2, good_lqual, 1, 10);
apply_power_test(lchan, PWR_TEST_RXLEV_TARGET_DBM - 2, good_lqual, 1, 9);
/* Check if BSC requesting a low max power is applied after loop calculation: */
lchan->ms_power_ctrl.max = ms_pwr_ctl_lvl(GSM_BAND_1800, 2);
OSMO_ASSERT(lchan->ms_power_ctrl.max == 14);
apply_power_test(lchan, PWR_TEST_RXLEV_TARGET_DBM + 2, good_lqual, 1, 14);
/* Set back a more normal max: */
lchan->ms_power_ctrl.max = ms_pwr_ctl_lvl(GSM_BAND_1800, 30);
OSMO_ASSERT(lchan->ms_power_ctrl.max == 0);
/* Disable dynamic power control and jump down */
lchan->ms_power_ctrl.dpc_params = NULL;
apply_power_test(lchan, -60, good_lqual, 0, 14);
/* Enable and leave it again */
lchan->ms_power_ctrl.dpc_params = &lchan->ms_dpc_params;
apply_power_test(lchan, -40, good_lqual, 1, 15);
}
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
static void test_pf_algo_ewma(void)
{
struct gsm_lchan *lchan;
const struct gsm_power_ctrl_params *params;
int16_t good_lqual;
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
const int *avg100;
init_test(__func__);
lchan = &g_trx->ts[0].lchan[0];
lchan->type = GSM_LCHAN_SDCCH;
params = lchan->ms_power_ctrl.dpc_params;
good_lqual = (params->ci_sdcch_meas.lower_thresh + 2) * 10;
avg100 = &lchan->ms_power_ctrl.rxlev_meas_proc.ewma.Avg100;
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
struct gsm_power_ctrl_meas_params *mp = &lchan->ms_dpc_params.rxlev_meas;
mp->algo = GSM_PWR_CTRL_MEAS_AVG_ALGO_OSMO_EWMA;
mp->ewma.alpha = 20; /* 80% smoothing */
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
lchan->ms_power_ctrl.current = ms_pwr_ctl_lvl(GSM_BAND_1800, 0);
OSMO_ASSERT(lchan->ms_power_ctrl.current == 15);
lchan->ms_power_ctrl.max = ms_pwr_ctl_lvl(GSM_BAND_1800, 26);
OSMO_ASSERT(lchan->ms_power_ctrl.max == 2);
#define CHECK_RXLEV_AVG100(exp) \
printf("\tAvg[t] is RxLev %2.2f (expected %2.2f)\n", \
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
((float) *avg100) / 100, exp);
/* UL RSSI remains constant => no UL power change */
apply_power_test(lchan, -75, good_lqual, 0, 15);
CHECK_RXLEV_AVG100((float)dbm2rxlev(-75)); /* RXLEV 35 */
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
/* Avg[t] = (0.2 * 20) + (0.8 * 35) = RXLEV 32, (-78 dBm) */
apply_power_test(lchan, -90, good_lqual, 1, 13); /* -90 dBm = RXLEV 20 */
CHECK_RXLEV_AVG100(32.00);
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
/* Avg[t] = (0.2 * 20) + (0.8 * 32) = RXLEV 29.6 (-80.4 dBm) */
apply_power_test(lchan, -90, good_lqual, 1, 11); /* -90 dBm = RXLEV 20 */
CHECK_RXLEV_AVG100(29.60);
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
/* Avg[t] = (0.2 * 40) + (0.8 * 29.60) = RXLEV 31.68 (-78.32 dBm),
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
* but due to up-/down-scaling artefacts we get the following:
* Avg100[t] = Avg100[t - 1] + A * (Pwr - Avg[t] / 100)
* Avg100[t] = 2960 + 20 * (40 - ((2960+50) / 100)) <- HERE we lose 0.1: (2960+50) / 100) = 30.1
* Avg100[t] = 2960 + 20 * (40 - 30) <- HERE we lose 20*0.1 = 2.0! (upscaled, hence we lose finally 2.0/100=0.2)
* Avg[t] = (3160) / 100 = 31.60*/
apply_power_test(lchan, -70, good_lqual, 1, 9); /* RXLEV 40 */
CHECK_RXLEV_AVG100(31.60);
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
mp->ewma.alpha = 70; /* 30% smoothing */
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
lchan->ms_power_ctrl.current = 15;
lchan->ms_power_ctrl.rxlev_meas_proc = \
(struct gsm_power_ctrl_meas_proc_state) { 0 };
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
/* This is the first sample, the filter outputs it as-is */
apply_power_test(lchan, -50, good_lqual, 0, 15); /* RXLEV 60 */
CHECK_RXLEV_AVG100((float)dbm2rxlev(-50));
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
/* Avg[t] = (0.7 * 60) + (0.3 * 60) = RXLEV 60 (-50.0 dBm) */
apply_power_test(lchan, -50, good_lqual, 0, 15);
CHECK_RXLEV_AVG100((float)dbm2rxlev(-50));
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
/* Simulate SACCH block loss (-110 dBm):
* Avg[t] = (0.7 * 0) + (0.3 * 60) = RXLEV 18.0 (-92.0 dBm) */
apply_power_test(lchan, -110, good_lqual, 1, 13); /* RXLEV 0 */
CHECK_RXLEV_AVG100(18.0);
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
}
static void test_power_hysteresis(void)
{
struct gsm_lchan *lchan;
const struct gsm_power_ctrl_params *params;
int16_t good_lqual;
init_test(__func__);
lchan = &g_trx->ts[0].lchan[0];
lchan->type = GSM_LCHAN_SDCCH;
params = lchan->ms_power_ctrl.dpc_params;
good_lqual = (params->ci_sdcch_meas.lower_thresh + 2) * 10;
/* Tolerate power deviations in range -80 .. -70 */
lchan->ms_dpc_params.rxlev_meas.lower_thresh = 30;
lchan->ms_dpc_params.rxlev_meas.upper_thresh = 40;
lchan->ms_power_ctrl.current = ms_pwr_ctl_lvl(GSM_BAND_1800, 0);
OSMO_ASSERT(lchan->ms_power_ctrl.current == 15);
lchan->ms_power_ctrl.max = ms_pwr_ctl_lvl(GSM_BAND_1800, 26);
OSMO_ASSERT(lchan->ms_power_ctrl.max == 2);
apply_power_test(lchan, PWR_TEST_RXLEV_TARGET_DBM, good_lqual, 0, 15);
apply_power_test(lchan, PWR_TEST_RXLEV_TARGET_DBM + 3, good_lqual, 0, 15);
apply_power_test(lchan, PWR_TEST_RXLEV_TARGET_DBM - 3, good_lqual, 0, 15);
apply_power_test(lchan, PWR_TEST_RXLEV_TARGET_DBM, good_lqual, 0, 15);
apply_power_test(lchan, PWR_TEST_RXLEV_TARGET_DBM + 5, good_lqual, 0, 15);
apply_power_test(lchan, PWR_TEST_RXLEV_TARGET_DBM - 5, good_lqual, 0, 15);
apply_power_test(lchan, PWR_TEST_RXLEV_TARGET_DBM - 10, good_lqual, 1, 13);
}
static void test_power_ctrl_interval(void)
{
struct gsm_lchan *lchan;
const struct gsm_power_ctrl_params *params;
int16_t good_lqual;
unsigned int i, j;
init_test(__func__);
lchan = &g_trx->ts[0].lchan[0];
lchan->type = GSM_LCHAN_SDCCH;
params = lchan->ms_power_ctrl.dpc_params;
good_lqual = (params->ci_sdcch_meas.lower_thresh + 2) * 10;
lchan->ms_power_ctrl.max = ms_pwr_ctl_lvl(GSM_BAND_1800, 26);
OSMO_ASSERT(lchan->ms_power_ctrl.max == 2);
const int script[][8][4] = {
{ /* P_Con_INTERVAL=0 (480 ms) */
/* { UL RxLev, expected rc, expected Tx power level } */
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 13 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 11 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 9 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 7 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 5 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 3 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 2 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 2 },
},
{ /* P_Con_INTERVAL=1 (960 ms) */
/* { UL RxLev, expected rc, expected Tx power level } */
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 13 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 0, 13 }, /* skipped */
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 11 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 0, 11 }, /* skipped */
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 9 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 0, 9 }, /* skipped */
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 7 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 0, 7 }, /* skipped */
},
{ /* P_Con_INTERVAL=2 (1920 ms) */
/* { UL RxLev, expected rc, expected Tx power level } */
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 13 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 0, 13 }, /* skipped */
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 0, 13 }, /* skipped */
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 0, 13 }, /* skipped */
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 1, 11 },
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 0, 11 }, /* skipped */
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 0, 11 }, /* skipped */
{ PWR_TEST_RXLEV_TARGET_DBM - 15, good_lqual, 0, 11 }, /* skipped */
},
};
for (i = 0; i < ARRAY_SIZE(script); i++) {
lchan->ms_power_ctrl.current = ms_pwr_ctl_lvl(GSM_BAND_1800, 0);
OSMO_ASSERT(lchan->ms_power_ctrl.current == 15);
/* Set the corresponding power control interval */
printf("%s(): power control interval is now %u\n", __func__, i);
lchan->ms_dpc_params.ctrl_interval = i;
for (j = 0; j < ARRAY_SIZE(script[i]); j++) {
apply_power_test(lchan, script[i][j][0], /* UL RxLev */
script[i][j][1], /* UL C/I */
script[i][j][2], /* expected rc */
script[i][j][3]); /* expected Tx power level */
}
printf("\n");
}
}
static void test_power_loop_ci(void)
{
struct gsm_lchan *lchan;
const struct gsm_power_ctrl_params *params;
int16_t good_lqual, too_low_lqual, too_high_lqual;
init_test(__func__);
lchan = &g_trx->ts[0].lchan[0];
params = lchan->ms_power_ctrl.dpc_params;
lchan->type = GSM_LCHAN_SDCCH;
good_lqual = (params->ci_sdcch_meas.lower_thresh + 2) * 10;
too_low_lqual = (params->ci_sdcch_meas.lower_thresh - 1) * 10;
too_high_lqual = (params->ci_sdcch_meas.upper_thresh + 1) * 10;
lchan->ms_power_ctrl.current = ms_pwr_ctl_lvl(GSM_BAND_1800, 0);
OSMO_ASSERT(lchan->ms_power_ctrl.current == 15);
lchan->ms_power_ctrl.max = ms_pwr_ctl_lvl(GSM_BAND_1800, 26);
OSMO_ASSERT(lchan->ms_power_ctrl.max == 2);
/* Simply clamping */
apply_power_test(lchan, -60, good_lqual, 0, 15);
/* Now UL C/I is too bad as well as RSSI: */
apply_power_test(lchan, -100, too_low_lqual, 1, 13);
apply_power_test(lchan, -100, too_low_lqual, 1, 11);
/* Now UL C/I is good again while RSSI is good: */
apply_power_test(lchan, -60, good_lqual, 1, 12);
apply_power_test(lchan, -60, too_high_lqual, 1, 13);
/* Now UL C/I is good while RSSI is bad, C/I mandates: */
apply_power_test(lchan, -100, good_lqual, 1, 11);
apply_power_test(lchan, -100, too_high_lqual, 1, 12);
/* Now UL C/I is bad again while RSSI is good, C/I mandates: */
apply_power_test(lchan, -60, good_lqual, 1, 13);
apply_power_test(lchan, -60, too_high_lqual, 1, 14);
}
/* Test whether ping pong between requested MS Power Level and announced MS
* Power level occurs, oscillating between considered good levels all the time:
* FIXME: Current code shows there's an issue with oscillating values. */
static void test_good_threshold_convergence(void)
{
struct gsm_lchan *lchan;
const struct gsm_power_ctrl_params *params;
int16_t good_lqual, good_rxlev;
init_test(__func__);
lchan = &g_trx->ts[0].lchan[0];
params = lchan->ms_power_ctrl.dpc_params;
lchan->ms_dpc_params.rxlev_meas.upper_thresh = 37;
lchan->ms_dpc_params.rxlev_meas.lower_thresh = 30;
lchan->type = GSM_LCHAN_SDCCH;
good_lqual = (params->ci_sdcch_meas.lower_thresh + 2) * 10;
good_rxlev = rxlev2dbm(params->rxlev_meas.lower_thresh + 2);
lchan->ms_power_ctrl.current = 10;
lchan->ms_power_ctrl.max = 2;
apply_power_test_ext(lchan, 9, good_rxlev, good_lqual, 0, 10);
apply_power_test_ext(lchan, 10, good_rxlev, good_lqual, 0, 10);
apply_power_test_ext(lchan, 9, good_rxlev, good_lqual, 0, 10);
apply_power_test_ext(lchan, 10, good_rxlev, good_lqual, 0, 10);
apply_power_test_ext(lchan, 9, good_rxlev, good_lqual, 0, 10);
}
int main(int argc, char **argv)
{
printf("Testing power loop...\n");
tall_bts_ctx = talloc_named_const(NULL, 1, "OsmoBTS context");
msgb_talloc_ctx_init(tall_bts_ctx, 0);
osmo_init_logging2(tall_bts_ctx, &bts_log_info);
osmo_stderr_target->categories[DLOOP].loglevel = LOGL_DEBUG;
osmo_stderr_target->categories[DL1C].loglevel = LOGL_DEBUG;
log_set_print_filename2(osmo_stderr_target, LOG_FILENAME_NONE);
log_set_use_color(osmo_stderr_target, 0);
log_set_print_category(osmo_stderr_target, 0);
log_set_print_category_hex(osmo_stderr_target, 0);
test_power_loop();
power_control: implement EWMA based Uplink power filtering So far the Uplink power control loop did not filter the Uplink RSSI measurements (reported by the BTS) at all. The lack of filtering makes our implementation too quick on the trigger, so in the real deployments there will be unneeded Tx power oscillations. In order to reduce this effect, let's implement a very simple EWMA (also known as Single Pole IIR) filtering that is defined as follows: Avg[n] = a * Pwr[n] + (1 - a) * Avg[n - 1] where parameter 'a' determines how much weight of the latest UL RSSI measurement result 'Pwr[n]' carries vs the weight of the average 'Avg[n - 1]'. The value of 'a' is usually a float in range 0 .. 1, so: - value 0.5 gives equal weight to both 'Pwr[n]' and 'Avg[n - 1]'; - value 1.0 means no filtering at all (pass through); - value 0.0 makes no sense. This formula was further optimized with the use of '+=' operator. The floating point math was also eliminated by scaling everything up (by 100). For more details, see: https://en.wikipedia.org/wiki/Moving_average https://en.wikipedia.org/wiki/Low-pass_filter#Simple_infinite_impulse_response_filter https://tomroelandts.com/articles/low-pass-single-pole-iir-filter The EWMA filtering is now *enabled by default*, but can be disabled or (re-)configured over the VTY at any time: ! Completely disable filtering no uplink-power-filtering ! Enable EWMA smoothing with the given parameters uplink-power-filtering algo ewma beta <1-99> Note that the VTY command expects 'beta' instead of 'alpha': alpha = (100 - beta) and the value must be in %. This is done for simplicity: 1% means lowest smoothing, 99% means highest smoothing. Let's say we have EWMA filtering enabled with alpha = 0.4, and get -98 dBm on the input, while the last output value was -60 dBm. The new output would be: Avg[n] = 0.4 * Pwr[n] + 0.6 * Avg[n - 1] Avg[n] = (0.4 * -98) + (0.6 * -60) Avg[n] = -75.2 => around -75 Of course, this is not a silver bullet, but better than nothing. Change-Id: Ib6dcadbf14ef59696c6a546bd323bda92d399f17 Related: SYS#4916
2020-10-13 14:40:24 +00:00
test_pf_algo_ewma();
test_power_hysteresis();
test_power_ctrl_interval();
test_power_loop_ci();
test_good_threshold_convergence();
printf("Power loop test OK\n");
return 0;
}