eliminate ATT steady-state error

classic Classic list List threaded Threaded
16 messages Options
Reply | Threaded
Open this post in threaded view
|

eliminate ATT steady-state error

Eltonwu
I use 5.2.1 ppz,and fly rotorcraft in ATT mode successfully,very stable.but have a small defect: always drift to left side.

from GCS, I see the rotorcraft's attitude has 2° degree error (not around 0 degree) when I release the attitude stick.so the rotorcraft drift to left.

I check the RC input found no attitude input.So I think it's caused by PID gains.

I try to increase the  ATT igain to 800,but found no improvement compare to 100 igain. So how does igain work?

How to eliminate ATT steady-state error?

Any suggestion will be appreciated.
Reply | Threaded
Open this post in threaded view
|

Re: eliminate ATT steady-state error

agressiva
try increase the i term.

btw, send you config file our pid gains

> Date: Wed, 20 May 2015 19:33:55 -0700

> From: [hidden email]
> To: [hidden email]
> Subject: [Paparazzi-devel] eliminate ATT steady-state error
>
> I use 5.2.1 ppz,and fly rotorcraft in ATT mode successfully,very stable.but
> have a small defect: always drift to left side.
>
> from GCS, I see the rotorcraft's attitude has 2° degree error (not around 0
> degree) when I release the attitude stick.so the rotorcraft drift to left.
>
> I check the RC input found no attitude input.So I think it's caused by PID
> gains.
>
> I try to increase the ATT igain to 800,but found no improvement compare to
> 100 igain. So how does igain work?
>
> How to eliminate ATT steady-state error?
>
> Any suggestion will be appreciated.
>
>
>
> --
> View this message in context: http://lists.paparazziuav.org/eliminate-ATT-steady-state-error-tp17019.html
> Sent from the paparazzi-devel mailing list archive at Nabble.com.
>
> _______________________________________________
> Paparazzi-devel mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: eliminate ATT steady-state error

Eltonwu
Hi Eduardo.

I have already tried to increase i term to 800 but found no improvement

quadrotor_wecam.xml

<airframe name="Quadrotor XYZ pwm">

  <modules main_freq="512">
    <load name="send_imu_mag_current.xml"/>
    <load name="openlog.xml"/>
    <load name="gps_ubx_ucenter.xml"/>
    <load name="sys_mon.xml"/>
    <load name="bat_checker.xml"/>
    <load name="geo_mag.xml"/>
     
  </modules>

  <firmware name="rotorcraft">
    <target name="ap" board="krooz_sd">
      <subsystem name="radio_control" type="datalink">
        <define name="RADIO_MODE" value="RADIO_AUX1"/>
        <configure name="USE_SECONDARY_SPEKTRUM_RECEIVER" value="0"/>
      </subsystem>
     
    </target>

    <target name="nps" board="pc">
      <subsystem name="fdm" type="jsbsim"/>
      <subsystem name="radio_control" type="ppm"/>
    </target>

    <subsystem name="motor_mixing"/>
    <subsystem name="actuators"     type="pwm">
      <define name="SERVO_HZ" value="400"/>
     
    </subsystem>

    <subsystem name="telemetry"     type="transparent"/>
    <subsystem name="imu"           type="krooz_sd"/>
    <subsystem name="gps"           type="ublox"/>
    <subsystem name="stabilization" type="int_quat"/>
    <subsystem name="ahrs"          type="int_cmpl_quat">
      <define name="AHRS_GRAVITY_HEURISTIC_FACTOR" value="30"/>
    </subsystem>
    <subsystem name="ins"           type="hff"/>

   
    <define name="USE_GPS_ACC4R" value="1"/>
    <define name="HFF_R_POS_MIN" value="2.5"/>
    <define name="HFF_R_SPEED_MIN" value="0.5"/>
    <define name="HFF_ACCEL_NOISE" value="1."/>
   
    <define name="VF_FLOAT_MEAS_NOISE" value="1."/>
  </firmware>

  <servos driver="Pwm">
    <servo name="NE"   no="0" min="1000" neutral="1150" max="1900"/>
    <servo name="SE"    no="1" min="1000" neutral="1150" max="1900"/>
    <servo name="SW"   no="2" min="1000" neutral="1150" max="1900"/>
    <servo name="NW"    no="3" min="1000" neutral="1150" max="1900"/>
  </servos>

  <commands>
    <axis name="ROLL"   failsafe_value="0"/>
    <axis name="PITCH"  failsafe_value="0"/>
    <axis name="YAW"    failsafe_value="0"/>
    <axis name="THRUST" failsafe_value="0"/>
  </commands>

  <command_laws>
    <call fun="motor_mixing_run(autopilot_motors_on,FALSE,values)"/>
    <set servo="NE" value="motor_mixing.commands[0]"/>
    <set servo="SE" value="motor_mixing.commands[1]"/>
    <set servo="SW" value="motor_mixing.commands[2]"/>
    <set servo="NW" value="motor_mixing.commands[3]"/>
  </command_laws>

  <section name="MIXING" prefix="MOTOR_MIXING_">
    <define name="TRIM_ROLL" value="0"/>
    <define name="TRIM_PITCH" value="0"/>
    <define name="TRIM_YAW" value="0"/>
    <define name="NB_MOTOR" value="4"/>
    <define name="SCALE" value="256"/>

    <define name="ROLL_COEF"   value="{ -MOTOR_MIXING_SCALE, -MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE }"/>
    <define name="PITCH_COEF"  value="{  MOTOR_MIXING_SCALE, -MOTOR_MIXING_SCALE, -MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE }"/>
    <define name="YAW_COEF"    value="{  MOTOR_MIXING_SCALE, -MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE, -MOTOR_MIXING_SCALE }"/>
    <define name="THRUST_COEF" value="{  MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE }"/>
  </section>

  <section name="IMU" prefix="IMU_">




<define name="ACCEL_X_NEUTRAL" value="116"/>
<define name="ACCEL_Y_NEUTRAL" value="128"/>
<define name="ACCEL_Z_NEUTRAL" value="-5957"/>
<define name="ACCEL_X_SENS" value="0.610012828874" integer="16"/>
<define name="ACCEL_Y_SENS" value="0.615211920657" integer="16"/>
<define name="ACCEL_Z_SENS" value="0.604669159053" integer="16"/>







   

   

   













<define name="MAG_X_NEUTRAL" value="63"/>
<define name="MAG_Y_NEUTRAL" value="17"/>
<define name="MAG_Z_NEUTRAL" value="-102"/>
<define name="MAG_X_SENS" value="4.22273667327" integer="16"/>
<define name="MAG_Y_SENS" value="4.32324776158" integer="16"/>
<define name="MAG_Z_SENS" value="4.56063086629" integer="16"/>





<define name= "MAG_X_CURRENT_COEF" value="-0.0234341805963"/>
<define name= "MAG_Y_CURRENT_COEF" value="0.0139393104786"/>
<define name= "MAG_Z_CURRENT_COEF" value="0.012462332939"/>




    <define name="BODY_TO_IMU_PHI"   value="0." unit="deg"/>
    <define name="BODY_TO_IMU_THETA" value="0." unit="deg"/>
    <define name="BODY_TO_IMU_PSI"   value="0." unit="deg"/>
  </section>

  <section name="AHRS" prefix="AHRS_">
    <define name="H_X" value="0.5531825"/>
    <define name="H_Y" value="0.1204294"/>
    <define name="H_Z" value="0.8243093"/>
  </section>

  <section name="INS" prefix="INS_">
  </section>

  <section name="STABILIZATION_RATE" prefix="STABILIZATION_RATE_">
   
    <define name="SP_MAX_P" value="10000"/>
    <define name="SP_MAX_Q" value="10000"/>
    <define name="SP_MAX_R" value="10000"/>
    <define name="DEADBAND_P" value="20"/>
    <define name="DEADBAND_Q" value="20"/>
    <define name="DEADBAND_R" value="200"/>
    <define name="REF_TAU" value="4"/>

   
    <define name="GAIN_P" value="400"/>
    <define name="GAIN_Q" value="400"/>
    <define name="GAIN_R" value="350"/>

    <define name="IGAIN_P" value="75"/>
    <define name="IGAIN_Q" value="75"/>
    <define name="IGAIN_R" value="50"/>

   
    <define name="DDGAIN_P" value="300"/>
    <define name="DDGAIN_Q" value="300"/>
    <define name="DDGAIN_R" value="300"/>
  </section>


  <section name="STABILIZATION_ATTITUDE" prefix="STABILIZATION_ATTITUDE_">

   
    <define name="SP_MAX_PHI" value="30." unit="deg"/>
    <define name="SP_MAX_THETA" value="30." unit="deg"/>
    <define name="SP_MAX_R" value="90." unit="deg/s"/>
    <define name="DEADBAND_R" value="250"/>
    <define name="DEADBAND_A" value="250"/>

   
    <define name="REF_OMEGA_P" value="400" unit="deg/s"/>
    <define name="REF_ZETA_P" value="0.85"/>
    <define name="REF_MAX_P" value="400." unit="deg/s"/>
    <define name="REF_MAX_PDOT" value="RadOfDeg(8000.)"/>

    <define name="REF_OMEGA_Q" value="400" unit="deg/s"/>
    <define name="REF_ZETA_Q" value="0.85"/>
    <define name="REF_MAX_Q" value="400." unit="deg/s"/>
    <define name="REF_MAX_QDOT" value="RadOfDeg(8000.)"/>

    <define name="REF_OMEGA_R" value="800" unit="deg/s"/>
    <define name="REF_ZETA_R" value="0.85"/>
    <define name="REF_MAX_R" value="90." unit="deg/s"/>
    <define name="REF_MAX_RDOT" value="RadOfDeg(1800.)"/>

   
    <define name="PHI_PGAIN" value="368"/>
    <define name="PHI_DGAIN" value="178"/>
    <define name="PHI_IGAIN" value="308"/>

    <define name="THETA_PGAIN" value="368"/>
    <define name="THETA_DGAIN" value="178"/>
    <define name="THETA_IGAIN" value="128"/>

   

    <define name="PSI_PGAIN" value="800"/>
    <define name="PSI_DGAIN" value="500"/>
    <define name="PSI_IGAIN" value="226"/>

   
    <define name="PHI_DDGAIN" value="90"/>
    <define name="THETA_DDGAIN" value="90"/>
    <define name="PSI_DDGAIN" value="400"/>

    <define name="PHI_AGAIN" value="0"/>
    <define name="THETA_AGAIN" value="0"/>
    <define name="PSI_AGAIN" value="0"/>
  </section>

  <section name="GUIDANCE_V" prefix="GUIDANCE_V_">
    <define name="MIN_ERR_Z"   value="POS_BFP_OF_REAL(-10.)"/>
    <define name="MAX_ERR_Z"   value="POS_BFP_OF_REAL( 10.)"/>
    <define name="MIN_ERR_ZD"  value="SPEED_BFP_OF_REAL(-10.)"/>
    <define name="MAX_ERR_ZD"  value="SPEED_BFP_OF_REAL( 10.)"/>
    <define name="MAX_SUM_ERR" value="2000000"/>
    <define name="REF_MIN_ZDD" value="-0.2*9.81"/> 
    <define name="REF_MAX_ZDD" value="0.2*9.81"/> 
    <define name="REF_MIN_ZD" value="-2."/>       
    <define name="REF_MAX_ZD" value="2."/>         
    <define name="HOVER_KP"    value="130"/>         
    <define name="HOVER_KD"    value="150"/>
    <define name="HOVER_KI"    value="10"/>
    <define name="ADAPT_NOISE_FACTOR" value="0.7"/>
    <define name="ADAPT_INITIAL_HOVER_THROTTLE" value="0.4"/>
    <define name="NOMINAL_HOVER_THROTTLE" value="0.9"/>
    <define name="ADAPT_THROTTLE_ENABLED" value="TRUE"/>
  </section>

  <section name="GUIDANCE_H" prefix="GUIDANCE_H_">
    <define name="MAX_BANK" value="27" unit="deg"/>
    <define name="USE_SPEED_REF" value="TRUE"/>
    <define name="REF_MAX_SPEED" value="2."/> 
    <define name="REF_MAX_ACCEL" value="7."/>
    <define name="PGAIN" value="68"/>  
    <define name="DGAIN" value="88"/> 
    <define name="VGAIN" value="0"/>
    <define name="AGAIN" value="88"/> 
    <define name="IGAIN" value="10"/> 
  </section>

  <section name="SIMULATOR" prefix="NPS_">
    <define name="ACTUATOR_NAMES"  value="{&quot;front_motor&quot;, &quot;back_motor&quot;, &quot;right_motor&quot;, &quot;left_motor&quot;}"/>
    <define name="JSBSIM_INIT" value="&quot;reset00&quot;"/>
    <define name="JSBSIM_MODEL" value="&quot;simple_quad&quot;"/>
    <define name="SENSORS_PARAMS" value="&quot;nps_sensors_params_default.h&quot;"/>
   
    <define name="JS_AXIS_MODE" value="4"/>
  </section>

  <section name="AUTOPILOT">
    <define name="MODE_MANUAL" value="AP_MODE_ATTITUDE_RC_CLIMB"/>
    <define name="MODE_AUTO1"  value="AP_MODE_HOVER_RC_CLIMB"/>
    <define name="MODE_AUTO2"  value="AP_MODE_NAV"/>
  </section>
Reply | Threaded
Open this post in threaded view
|

Re: eliminate ATT steady-state error

Sergey Krukowski
I found that there is any improvement at all is increasing igain to about  
2000-8000. (I'm using euler stabilization however). The probable reason  
for that in quat stabilization is that the IERROR_SCALE value is too high:
https://github.com/paparazzi/paparazzi/blob/master/sw/airborne/firmwares/rotorcraft/stabilization/stabilization_attitude_quat_int.c#L255-L257

Best regards,
Sergey

> Hi Eduardo.
>
> I have already tried to increase i term to 800 but found no improvement
>
> quadrotor_wecam.xml
> <http://lists.paparazziuav.org/file/n17022/quadrotor_wecam.xml>
>
> <airframe name="Quadrotor XYZ pwm">
>
>   <modules main_freq="512">
>     <load name="send_imu_mag_current.xml"/>
>     <load name="openlog.xml"/>
>     <load name="gps_ubx_ucenter.xml"/>
>     <load name="sys_mon.xml"/>
>     <load name="bat_checker.xml"/>
>     <load name="geo_mag.xml"/>
>  </modules>
>
>   <firmware name="rotorcraft">
>     <target name="ap" board="krooz_sd">
>       <subsystem name="radio_control" type="datalink">
>         <define name="RADIO_MODE" value="RADIO_AUX1"/>
>         <configure name="USE_SECONDARY_SPEKTRUM_RECEIVER" value="0"/>
>       </subsystem>
>    </target>
>
>     <target name="nps" board="pc">
>       <subsystem name="fdm" type="jsbsim"/>
>       <subsystem name="radio_control" type="ppm"/>
>     </target>
>
>     <subsystem name="motor_mixing"/>
>     <subsystem name="actuators"     type="pwm">
>       <define name="SERVO_HZ" value="400"/>
>    </subsystem>
>
>     <subsystem name="telemetry"     type="transparent"/>
>     <subsystem name="imu"           type="krooz_sd"/>
>     <subsystem name="gps"           type="ublox"/>
>     <subsystem name="stabilization" type="int_quat"/>
>     <subsystem name="ahrs"          type="int_cmpl_quat">
>       <define name="AHRS_GRAVITY_HEURISTIC_FACTOR" value="30"/>
>     </subsystem>
>     <subsystem name="ins"           type="hff"/>
>
>    <define name="USE_GPS_ACC4R" value="1"/>
>     <define name="HFF_R_POS_MIN" value="2.5"/>
>     <define name="HFF_R_SPEED_MIN" value="0.5"/>
>     <define name="HFF_ACCEL_NOISE" value="1."/>
>    <define name="VF_FLOAT_MEAS_NOISE" value="1."/>
>   </firmware>
>
>   <servos driver="Pwm">
>     <servo name="NE"   no="0" min="1000" neutral="1150" max="1900"/>
>     <servo name="SE"    no="1" min="1000" neutral="1150" max="1900"/>
>     <servo name="SW"   no="2" min="1000" neutral="1150" max="1900"/>
>     <servo name="NW"    no="3" min="1000" neutral="1150" max="1900"/>
>   </servos>
>
>   <commands>
>     <axis name="ROLL"   failsafe_value="0"/>
>     <axis name="PITCH"  failsafe_value="0"/>
>     <axis name="YAW"    failsafe_value="0"/>
>     <axis name="THRUST" failsafe_value="0"/>
>   </commands>
>
>   <command_laws>
>     <call fun="motor_mixing_run(autopilot_motors_on,FALSE,values)"/>
>     <set servo="NE" value="motor_mixing.commands[0]"/>
>     <set servo="SE" value="motor_mixing.commands[1]"/>
>     <set servo="SW" value="motor_mixing.commands[2]"/>
>     <set servo="NW" value="motor_mixing.commands[3]"/>
>   </command_laws>
>
>   <section name="MIXING" prefix="MOTOR_MIXING_">
>     <define name="TRIM_ROLL" value="0"/>
>     <define name="TRIM_PITCH" value="0"/>
>     <define name="TRIM_YAW" value="0"/>
>     <define name="NB_MOTOR" value="4"/>
>     <define name="SCALE" value="256"/>
>
>     <define name="ROLL_COEF"   value="{ -MOTOR_MIXING_SCALE,
> -MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE }"/>
>     <define name="PITCH_COEF"  value="{  MOTOR_MIXING_SCALE,
> -MOTOR_MIXING_SCALE, -MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE }"/>
>     <define name="YAW_COEF"    value="{  MOTOR_MIXING_SCALE,
> -MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE, -MOTOR_MIXING_SCALE }"/>
>     <define name="THRUST_COEF" value="{  MOTOR_MIXING_SCALE,
> MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE,  MOTOR_MIXING_SCALE }"/>
>   </section>
>
>   <section name="IMU" prefix="IMU_">
>
>
>


--
Erstellt mit Operas revolutionärem E-Mail-Modul: http://www.opera.com/mail/

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: eliminate ATT steady-state error

agressiva
Sergey, i reduce the IERROR_SCALE to 256 and my quad fly very well.

> To: [hidden email]

> Date: Fri, 22 May 2015 08:00:55 +0200
> From: [hidden email]
> Subject: Re: [Paparazzi-devel] eliminate ATT steady-state error
>
> I found that there is any improvement at all is increasing igain to about
> 2000-8000. (I'm using euler stabilization however). The probable reason
> for that in quat stabilization is that the IERROR_SCALE value is too high:
> https://github.com/paparazzi/paparazzi/blob/master/sw/airborne/firmwares/rotorcraft/stabilization/stabilization_attitude_quat_int.c#L255-L257
>
> Best regards,
> Sergey
>
> > Hi Eduardo.
> >
> > I have already tried to increase i term to 800 but found no improvement
> >
> > quadrotor_wecam.xml
> > <http://lists.paparazziuav.org/file/n17022/quadrotor_wecam.xml>
> >
> > <airframe name="Quadrotor XYZ pwm">
> >
> > <modules main_freq="512">
> > <load name="send_imu_mag_current.xml"/>
> > <load name="openlog.xml"/>
> > <load name="gps_ubx_ucenter.xml"/>
> > <load name="sys_mon.xml"/>
> > <load name="bat_checker.xml"/>
> > <load name="geo_mag.xml"/>
> > </modules>
> >
> > <firmware name="rotorcraft">
> > <target name="ap" board="krooz_sd">
> > <subsystem name="radio_control" type="datalink">
> > <define name="RADIO_MODE" value="RADIO_AUX1"/>
> > <configure name="USE_SECONDARY_SPEKTRUM_RECEIVER" value="0"/>
> > </subsystem>
> > </target>
> >
> > <target name="nps" board="pc">
> > <subsystem name="fdm" type="jsbsim"/>
> > <subsystem name="radio_control" type="ppm"/>
> > </target>
> >
> > <subsystem name="motor_mixing"/>
> > <subsystem name="actuators" type="pwm">
> > <define name="SERVO_HZ" value="400"/>
> > </subsystem>
> >
> > <subsystem name="telemetry" type="transparent"/>
> > <subsystem name="imu" type="krooz_sd"/>
> > <subsystem name="gps" type="ublox"/>
> > <subsystem name="stabilization" type="int_quat"/>
> > <subsystem name="ahrs" type="int_cmpl_quat">
> > <define name="AHRS_GRAVITY_HEURISTIC_FACTOR" value="30"/>
> > </subsystem>
> > <subsystem name="ins" type="hff"/>
> >
> > <define name="USE_GPS_ACC4R" value="1"/>
> > <define name="HFF_R_POS_MIN" value="2.5"/>
> > <define name="HFF_R_SPEED_MIN" value="0.5"/>
> > <define name="HFF_ACCEL_NOISE" value="1."/>
> > <define name="VF_FLOAT_MEAS_NOISE" value="1."/>
> > </firmware>
> >
> > <servos driver="Pwm">
> > <servo name="NE" no="0" min="1000" neutral="1150" max="1900"/>
> > <servo name="SE" no="1" min="1000" neutral="1150" max="1900"/>
> > <servo name="SW" no="2" min="1000" neutral="1150" max="1900"/>
> > <servo name="NW" no="3" min="1000" neutral="1150" max="1900"/>
> > </servos>
> >
> > <commands>
> > <axis name="ROLL" failsafe_value="0"/>
> > <axis name="PITCH" failsafe_value="0"/>
> > <axis name="YAW" failsafe_value="0"/>
> > <axis name="THRUST" failsafe_value="0"/>
> > </commands>
> >
> > <command_laws>
> > <call fun="motor_mixing_run(autopilot_motors_on,FALSE,values)"/>
> > <set servo="NE" value="motor_mixing.commands[0]"/>
> > <set servo="SE" value="motor_mixing.commands[1]"/>
> > <set servo="SW" value="motor_mixing.commands[2]"/>
> > <set servo="NW" value="motor_mixing.commands[3]"/>
> > </command_laws>
> >
> > <section name="MIXING" prefix="MOTOR_MIXING_">
> > <define name="TRIM_ROLL" value="0"/>
> > <define name="TRIM_PITCH" value="0"/>
> > <define name="TRIM_YAW" value="0"/>
> > <define name="NB_MOTOR" value="4"/>
> > <define name="SCALE" value="256"/>
> >
> > <define name="ROLL_COEF" value="{ -MOTOR_MIXING_SCALE,
> > -MOTOR_MIXING_SCALE, MOTOR_MIXING_SCALE, MOTOR_MIXING_SCALE }"/>
> > <define name="PITCH_COEF" value="{ MOTOR_MIXING_SCALE,
> > -MOTOR_MIXING_SCALE, -MOTOR_MIXING_SCALE, MOTOR_MIXING_SCALE }"/>
> > <define name="YAW_COEF" value="{ MOTOR_MIXING_SCALE,
> > -MOTOR_MIXING_SCALE, MOTOR_MIXING_SCALE, -MOTOR_MIXING_SCALE }"/>
> > <define name="THRUST_COEF" value="{ MOTOR_MIXING_SCALE,
> > MOTOR_MIXING_SCALE, MOTOR_MIXING_SCALE, MOTOR_MIXING_SCALE }"/>
> > </section>
> >
> > <section name="IMU" prefix="IMU_">
> >
> >
> >
>
>
> --
> Erstellt mit Operas revolutionärem E-Mail-Modul: http://www.opera.com/mail/
>
> _______________________________________________
> Paparazzi-devel mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: eliminate ATT steady-state error

Eltonwu
Hi all
ISCALE_ERROR reduce to 128 works perfect

Thank you very much
Reply | Threaded
Open this post in threaded view
|

Lisa/M terminated standby circling

Balázs GÁTI
Hi,

I just experienced an interesting behavior of a Lisa/M v2 (Transition
Robotics) with Paparazzi v5.4.0.

I had finished tuning my fixed wing in AUTO1 and switched to AUTO2. The
plane started to intercept altitude and to align himself to the Standby
circle. After a half (more or less) stabilized circle she terminated the
circle and started to roll to the opposite direction. I switched back to
AUTO1 and landed. Post flight analysis of log showed no anomaly. No
warning, no lost RC or GPS, no GPWS warning, no change in PPRZMode,
motor running.

I increased the height of waypoints and started a new flight. The
result was the same. After a half circle in Standby mode, she started an
opposite spiral. I left her to do it and experienced a path similar to
one and a half spiral in strong in wind by constant roll. Post flight
analysis showed a rapid change in Desired.Roll, even the carrot and the
green circle of the Standby mode still behaved as I expected.

The rapid change in Desired.Roll happened more or less in the same time
as Attitude.Psi changed sign. This is only expectation, because this
sign error would ruin all Paparazzi flight in the world. Even my
simulations didn't predict this behavior. Unfortunately the telemetry
messages are not too dense to confirm this expectation. Do you have any
advice what should I do to identify the reasons? (NEO7P works already
fine.)


Regards,
  Balazs

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: Lisa/M terminated standby circling

Gautier Hattenberger-3
Can you post the log files somewhere ? And give exact paparzzi version
(./paparazzi_version)

Gautier

Le 29/05/2015 15:22, Balázs GÁTI a écrit :

> Hi,
>
> I just experienced an interesting behavior of a Lisa/M v2 (Transition
> Robotics) with Paparazzi v5.4.0.
>
> I had finished tuning my fixed wing in AUTO1 and switched to AUTO2.
> The plane started to intercept altitude and to align himself to the
> Standby circle. After a half (more or less) stabilized circle she
> terminated the circle and started to roll to the opposite direction. I
> switched back to AUTO1 and landed. Post flight analysis of log showed
> no anomaly. No warning, no lost RC or GPS, no GPWS warning, no change
> in PPRZMode, motor running.
>
> I increased the height of waypoints and started a new flight. The
> result was the same. After a half circle in Standby mode, she started
> an opposite spiral. I left her to do it and experienced a path similar
> to one and a half spiral in strong in wind by constant roll. Post
> flight analysis showed a rapid change in Desired.Roll, even the carrot
> and the green circle of the Standby mode still behaved as I expected.
>
> The rapid change in Desired.Roll happened more or less in the same
> time as Attitude.Psi changed sign. This is only expectation, because
> this sign error would ruin all Paparazzi flight in the world. Even my
> simulations didn't predict this behavior. Unfortunately the telemetry
> messages are not too dense to confirm this expectation. Do you have
> any advice what should I do to identify the reasons? (NEO7P works
> already fine.)
>
>
> Regards,
>  Balazs
>
> _______________________________________________
> Paparazzi-devel mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel


_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: Lisa/M terminated standby circling

Balázs GÁTI
Hi,

Paparazzi version: v5.4.0_stable

The two log files can be downloaded:

ftp.edufly.net Port 21
user: [hidden email]
p: NW7100log

Time of interest
13_03_52: t= 315s
13_47_23: t= 320s


Balazs



2015-05-29 15:43 időpontban Gautier Hattenberger ezt írta:

> Can you post the log files somewhere ? And give exact paparzzi
> version (./paparazzi_version)
>
> Gautier
>
> Le 29/05/2015 15:22, Balázs GÁTI a écrit :
>> Hi,
>>
>> I just experienced an interesting behavior of a Lisa/M v2
>> (Transition Robotics) with Paparazzi v5.4.0.
>>
>> I had finished tuning my fixed wing in AUTO1 and switched to AUTO2.
>> The plane started to intercept altitude and to align himself to the
>> Standby circle. After a half (more or less) stabilized circle she
>> terminated the circle and started to roll to the opposite direction. I
>> switched back to AUTO1 and landed. Post flight analysis of log showed
>> no anomaly. No warning, no lost RC or GPS, no GPWS warning, no change
>> in PPRZMode, motor running.
>>
>> I increased the height of waypoints and started a new flight. The
>> result was the same. After a half circle in Standby mode, she started
>> an opposite spiral. I left her to do it and experienced a path similar
>> to one and a half spiral in strong in wind by constant roll. Post
>> flight analysis showed a rapid change in Desired.Roll, even the carrot
>> and the green circle of the Standby mode still behaved as I expected.
>>
>> The rapid change in Desired.Roll happened more or less in the same
>> time as Attitude.Psi changed sign. This is only expectation, because
>> this sign error would ruin all Paparazzi flight in the world. Even my
>> simulations didn't predict this behavior. Unfortunately the telemetry
>> messages are not too dense to confirm this expectation. Do you have
>> any advice what should I do to identify the reasons? (NEO7P works
>> already fine.)
>>
>>
>> Regards,
>>  Balazs
>>
>> _______________________________________________
>> Paparazzi-devel mailing list
>> [hidden email]
>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>
>
> _______________________________________________
> Paparazzi-devel mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

---
Balazs GATI, PhD
associate professor
Department of Aeronautics, Naval Architecture and Railway
Vehicles
Budapest University of Technology and Economics

Address:   Budapest
Stoczek u 6. J. ép. 423
1111
Tel:       +(36)-1-463-1960 [1]
Fax:       +(36)-1-463-3080 [2]
Homepage: http://vrht.bme.hu/ [3]

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: Lisa/M terminated standby circling

Gautier Hattenberger-3
It will not solve all the problems, but I think you have not activated
the VELNED message in your u-blox module. It is annoying since on
fixed-wing, it is used to get the (ground) speed and direction. It is
even a miracle that you could fly in AUTO2.

The strange trajectory occurs when the heading is wrapping at 180 deg
while the desired heading is wrapping at 360. We have an open issue for
this... I have never used the int_cmpl_quat ahrs on FW, especially with
MAG instead of GPS course.

If activating the VELNED message is not enough, we'll have to take a
closer look at how the heading error is computed in nav.

Gautier


Le 29/05/2015 16:11, Balázs GÁTI a écrit :

> Hi,
>
> Paparazzi version: v5.4.0_stable
>
> The two log files can be downloaded:
>
> ftp.edufly.net Port 21
> user: [hidden email]
> p: NW7100log
>
> Time of interest
> 13_03_52: t= 315s
> 13_47_23: t= 320s
>
>
> Balazs
>
>
>
> 2015-05-29 15:43 időpontban Gautier Hattenberger ezt írta:
>> Can you post the log files somewhere ? And give exact paparzzi
>> version (./paparazzi_version)
>>
>> Gautier
>>
>> Le 29/05/2015 15:22, Balázs GÁTI a écrit :
>>> Hi,
>>>
>>> I just experienced an interesting behavior of a Lisa/M v2
>>> (Transition Robotics) with Paparazzi v5.4.0.
>>>
>>> I had finished tuning my fixed wing in AUTO1 and switched to AUTO2.
>>> The plane started to intercept altitude and to align himself to the
>>> Standby circle. After a half (more or less) stabilized circle she
>>> terminated the circle and started to roll to the opposite direction.
>>> I switched back to AUTO1 and landed. Post flight analysis of log
>>> showed no anomaly. No warning, no lost RC or GPS, no GPWS warning,
>>> no change in PPRZMode, motor running.
>>>
>>> I increased the height of waypoints and started a new flight. The
>>> result was the same. After a half circle in Standby mode, she
>>> started an opposite spiral. I left her to do it and experienced a
>>> path similar to one and a half spiral in strong in wind by constant
>>> roll. Post flight analysis showed a rapid change in Desired.Roll,
>>> even the carrot and the green circle of the Standby mode still
>>> behaved as I expected.
>>>
>>> The rapid change in Desired.Roll happened more or less in the same
>>> time as Attitude.Psi changed sign. This is only expectation, because
>>> this sign error would ruin all Paparazzi flight in the world. Even
>>> my simulations didn't predict this behavior. Unfortunately the
>>> telemetry messages are not too dense to confirm this expectation. Do
>>> you have any advice what should I do to identify the reasons? (NEO7P
>>> works already fine.)
>>>
>>>
>>> Regards,
>>>  Balazs
>>>
>>> _______________________________________________
>>> Paparazzi-devel mailing list
>>> [hidden email]
>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>
>>
>> _______________________________________________
>> Paparazzi-devel mailing list
>> [hidden email]
>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>
> ---
> Balazs GATI, PhD
> associate professor
> Department of Aeronautics, Naval Architecture and Railway
> Vehicles
> Budapest University of Technology and Economics
>
> Address:   Budapest
> Stoczek u 6. J. ép. 423
> 1111
> Tel:       +(36)-1-463-1960 [1]
> Fax:       +(36)-1-463-3080 [2]
> Homepage: http://vrht.bme.hu/ [3]
>
> _______________________________________________
> Paparazzi-devel mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel


_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: Lisa/M terminated standby circling

Balázs GÁTI
Hi,

I am very pleased for the advice and useful informations! My airframe
file is a result of trial-and-error procedure, and perhaps contains some
never tested combination of settings. I have one more suspect: the
BODY_TO_IMU_PSI = 180 because the opposit attitude of the Lisa. I'll
change it to 179.9

I am planning the next flight. I'll switch VELNED on. Can I do
something more to support the debugging?

Regards,
  Balazs


2015-05-29 18:14 időpontban Gautier Hattenberger ezt írta:

> It will not solve all the problems, but I think you have not
> activated the VELNED message in your u-blox module. It is annoying
> since on fixed-wing, it is used to get the (ground) speed and
> direction. It is even a miracle that you could fly in AUTO2.
>
> The strange trajectory occurs when the heading is wrapping at 180 deg
> while the desired heading is wrapping at 360. We have an open issue
> for this... I have never used the int_cmpl_quat ahrs on FW,
> especially
> with MAG instead of GPS course.
>
> If activating the VELNED message is not enough, we'll have to take a
> closer look at how the heading error is computed in nav.
>
> Gautier
>
>
> Le 29/05/2015 16:11, Balázs GÁTI a écrit :
>> Hi,
>>
>> Paparazzi version: v5.4.0_stable
>>
>> The two log files can be downloaded:
>>
>> ftp.edufly.net Port 21
>> user: [hidden email]
>> p: NW7100log
>>
>> Time of interest
>> 13_03_52: t= 315s
>> 13_47_23: t= 320s
>>
>>
>> Balazs
>>
>>
>>
>> 2015-05-29 15:43 időpontban Gautier Hattenberger ezt írta:
>>> Can you post the log files somewhere ? And give exact paparzzi
>>> version (./paparazzi_version)
>>>
>>> Gautier
>>>
>>> Le 29/05/2015 15:22, Balázs GÁTI a écrit :
>>>> Hi,
>>>>
>>>> I just experienced an interesting behavior of a Lisa/M v2
>>>> (Transition Robotics) with Paparazzi v5.4.0.
>>>>
>>>> I had finished tuning my fixed wing in AUTO1 and switched to
>>>> AUTO2. The plane started to intercept altitude and to align himself
>>>> to the Standby circle. After a half (more or less) stabilized circle
>>>> she terminated the circle and started to roll to the opposite
>>>> direction. I switched back to AUTO1 and landed. Post flight analysis
>>>> of log showed no anomaly. No warning, no lost RC or GPS, no GPWS
>>>> warning, no change in PPRZMode, motor running.
>>>>
>>>> I increased the height of waypoints and started a new flight. The
>>>> result was the same. After a half circle in Standby mode, she
>>>> started an opposite spiral. I left her to do it and experienced a
>>>> path similar to one and a half spiral in strong in wind by constant
>>>> roll. Post flight analysis showed a rapid change in Desired.Roll,
>>>> even the carrot and the green circle of the Standby mode still
>>>> behaved as I expected.
>>>>
>>>> The rapid change in Desired.Roll happened more or less in the same
>>>> time as Attitude.Psi changed sign. This is only expectation, because
>>>> this sign error would ruin all Paparazzi flight in the world. Even
>>>> my simulations didn't predict this behavior. Unfortunately the
>>>> telemetry messages are not too dense to confirm this expectation. Do
>>>> you have any advice what should I do to identify the reasons? (NEO7P
>>>> works already fine.)
>>>>
>>>>
>>>> Regards,
>>>>  Balazs
>>>>
>>>> _______________________________________________
>>>> Paparazzi-devel mailing list
>>>> [hidden email]
>>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>>
>>>
>>> _______________________________________________
>>> Paparazzi-devel mailing list
>>> [hidden email]
>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>
>> ---
>> Balazs GATI, PhD
>> associate professor
>> Department of Aeronautics, Naval Architecture and Railway
>> Vehicles
>> Budapest University of Technology and Economics
>>
>> Address:   Budapest
>> Stoczek u 6. J. ép. 423
>> 1111
>> Tel:       +(36)-1-463-1960 [1]
>> Fax:       +(36)-1-463-3080 [2]
>> Homepage: http://vrht.bme.hu/ [3]
>>
>> _______________________________________________
>> Paparazzi-devel mailing list
>> [hidden email]
>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>
>
> _______________________________________________
> Paparazzi-devel mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

---
Balazs GATI, PhD
associate professor
Department of Aeronautics, Naval Architecture and Railway
Vehicles
Budapest University of Technology and Economics

Address:   Budapest
Stoczek u 6. J. ép. 423
1111
Tel:       +(36)-1-463-1960 [1]
Fax:       +(36)-1-463-3080 [2]
Homepage: http://vrht.bme.hu/ [3]

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: Lisa/M terminated standby circling

Gautier Hattenberger-3
In reply to this post by Balázs GÁTI
After looking at the nav code, your problem really seems to be the
missing VELNED message in your GPS module conf.

G.


Le 29/05/2015 16:11, Balázs GÁTI a écrit :

> Hi,
>
> Paparazzi version: v5.4.0_stable
>
> The two log files can be downloaded:
>
> ftp.edufly.net Port 21
> user: [hidden email]
> p: NW7100log
>
> Time of interest
> 13_03_52: t= 315s
> 13_47_23: t= 320s
>
>
> Balazs
>
>
>
> 2015-05-29 15:43 időpontban Gautier Hattenberger ezt írta:
>> Can you post the log files somewhere ? And give exact paparzzi
>> version (./paparazzi_version)
>>
>> Gautier
>>
>> Le 29/05/2015 15:22, Balázs GÁTI a écrit :
>>> Hi,
>>>
>>> I just experienced an interesting behavior of a Lisa/M v2
>>> (Transition Robotics) with Paparazzi v5.4.0.
>>>
>>> I had finished tuning my fixed wing in AUTO1 and switched to AUTO2.
>>> The plane started to intercept altitude and to align himself to the
>>> Standby circle. After a half (more or less) stabilized circle she
>>> terminated the circle and started to roll to the opposite direction.
>>> I switched back to AUTO1 and landed. Post flight analysis of log
>>> showed no anomaly. No warning, no lost RC or GPS, no GPWS warning,
>>> no change in PPRZMode, motor running.
>>>
>>> I increased the height of waypoints and started a new flight. The
>>> result was the same. After a half circle in Standby mode, she
>>> started an opposite spiral. I left her to do it and experienced a
>>> path similar to one and a half spiral in strong in wind by constant
>>> roll. Post flight analysis showed a rapid change in Desired.Roll,
>>> even the carrot and the green circle of the Standby mode still
>>> behaved as I expected.
>>>
>>> The rapid change in Desired.Roll happened more or less in the same
>>> time as Attitude.Psi changed sign. This is only expectation, because
>>> this sign error would ruin all Paparazzi flight in the world. Even
>>> my simulations didn't predict this behavior. Unfortunately the
>>> telemetry messages are not too dense to confirm this expectation. Do
>>> you have any advice what should I do to identify the reasons? (NEO7P
>>> works already fine.)
>>>
>>>
>>> Regards,
>>>  Balazs
>>>
>>> _______________________________________________
>>> Paparazzi-devel mailing list
>>> [hidden email]
>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>
>>
>> _______________________________________________
>> Paparazzi-devel mailing list
>> [hidden email]
>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>
> ---
> Balazs GATI, PhD
> associate professor
> Department of Aeronautics, Naval Architecture and Railway
> Vehicles
> Budapest University of Technology and Economics
>
> Address:   Budapest
> Stoczek u 6. J. ép. 423
> 1111
> Tel:       +(36)-1-463-1960 [1]
> Fax:       +(36)-1-463-3080 [2]
> Homepage: http://vrht.bme.hu/ [3]
>
> _______________________________________________
> Paparazzi-devel mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel


_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: Lisa/M terminated standby circling

Gautier Hattenberger-3
In reply to this post by Balázs GÁTI
Reporting issues with accurate description and information is already good!
After that, it mostly depends on your programming skills (and nerves to
look at the code in depth...)


Le 29/05/2015 18:37, Balázs GÁTI a écrit :

> Hi,
>
> I am very pleased for the advice and useful informations! My airframe
> file is a result of trial-and-error procedure, and perhaps contains
> some never tested combination of settings. I have one more suspect:
> the BODY_TO_IMU_PSI = 180 because the opposit attitude of the Lisa.
> I'll change it to 179.9
>
> I am planning the next flight. I'll switch VELNED on. Can I do
> something more to support the debugging?
>
> Regards,
>  Balazs
>
>
> 2015-05-29 18:14 időpontban Gautier Hattenberger ezt írta:
>> It will not solve all the problems, but I think you have not
>> activated the VELNED message in your u-blox module. It is annoying
>> since on fixed-wing, it is used to get the (ground) speed and
>> direction. It is even a miracle that you could fly in AUTO2.
>>
>> The strange trajectory occurs when the heading is wrapping at 180 deg
>> while the desired heading is wrapping at 360. We have an open issue
>> for this... I have never used the int_cmpl_quat ahrs on FW, especially
>> with MAG instead of GPS course.
>>
>> If activating the VELNED message is not enough, we'll have to take a
>> closer look at how the heading error is computed in nav.
>>
>> Gautier
>>
>>
>> Le 29/05/2015 16:11, Balázs GÁTI a écrit :
>>> Hi,
>>>
>>> Paparazzi version: v5.4.0_stable
>>>
>>> The two log files can be downloaded:
>>>
>>> ftp.edufly.net Port 21
>>> user: [hidden email]
>>> p: NW7100log
>>>
>>> Time of interest
>>> 13_03_52: t= 315s
>>> 13_47_23: t= 320s
>>>
>>>
>>> Balazs
>>>
>>>
>>>
>>> 2015-05-29 15:43 időpontban Gautier Hattenberger ezt írta:
>>>> Can you post the log files somewhere ? And give exact paparzzi
>>>> version (./paparazzi_version)
>>>>
>>>> Gautier
>>>>
>>>> Le 29/05/2015 15:22, Balázs GÁTI a écrit :
>>>>> Hi,
>>>>>
>>>>> I just experienced an interesting behavior of a Lisa/M v2
>>>>> (Transition Robotics) with Paparazzi v5.4.0.
>>>>>
>>>>> I had finished tuning my fixed wing in AUTO1 and switched to
>>>>> AUTO2. The plane started to intercept altitude and to align
>>>>> himself to the Standby circle. After a half (more or less)
>>>>> stabilized circle she terminated the circle and started to roll to
>>>>> the opposite direction. I switched back to AUTO1 and landed. Post
>>>>> flight analysis of log showed no anomaly. No warning, no lost RC
>>>>> or GPS, no GPWS warning, no change in PPRZMode, motor running.
>>>>>
>>>>> I increased the height of waypoints and started a new flight. The
>>>>> result was the same. After a half circle in Standby mode, she
>>>>> started an opposite spiral. I left her to do it and experienced a
>>>>> path similar to one and a half spiral in strong in wind by
>>>>> constant roll. Post flight analysis showed a rapid change in
>>>>> Desired.Roll, even the carrot and the green circle of the Standby
>>>>> mode still behaved as I expected.
>>>>>
>>>>> The rapid change in Desired.Roll happened more or less in the same
>>>>> time as Attitude.Psi changed sign. This is only expectation,
>>>>> because this sign error would ruin all Paparazzi flight in the
>>>>> world. Even my simulations didn't predict this behavior.
>>>>> Unfortunately the telemetry messages are not too dense to confirm
>>>>> this expectation. Do you have any advice what should I do to
>>>>> identify the reasons? (NEO7P works already fine.)
>>>>>
>>>>>
>>>>> Regards,
>>>>>  Balazs
>>>>>
>>>>> _______________________________________________
>>>>> Paparazzi-devel mailing list
>>>>> [hidden email]
>>>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>>>
>>>>
>>>> _______________________________________________
>>>> Paparazzi-devel mailing list
>>>> [hidden email]
>>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>>
>>> ---
>>> Balazs GATI, PhD
>>> associate professor
>>> Department of Aeronautics, Naval Architecture and Railway
>>> Vehicles
>>> Budapest University of Technology and Economics
>>>
>>> Address:   Budapest
>>> Stoczek u 6. J. ép. 423
>>> 1111
>>> Tel:       +(36)-1-463-1960 [1]
>>> Fax:       +(36)-1-463-3080 [2]
>>> Homepage: http://vrht.bme.hu/ [3]
>>>
>>> _______________________________________________
>>> Paparazzi-devel mailing list
>>> [hidden email]
>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>
>>
>> _______________________________________________
>> Paparazzi-devel mailing list
>> [hidden email]
>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>
> ---
> Balazs GATI, PhD
> associate professor
> Department of Aeronautics, Naval Architecture and Railway
> Vehicles
> Budapest University of Technology and Economics
>
> Address:   Budapest
> Stoczek u 6. J. ép. 423
> 1111
> Tel:       +(36)-1-463-1960 [1]
> Fax:       +(36)-1-463-3080 [2]
> Homepage: http://vrht.bme.hu/ [3]
>
> _______________________________________________
> Paparazzi-devel mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel


_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: Lisa/M terminated standby circling [SOLVED]

Balázs GÁTI
In reply to this post by Gautier Hattenberger-3
Hi,

I've implemented the GPS Ublox Ucenter module in the airframe file
(http://wiki.paparazziuav.org/wiki/Module/GPS_UBlox_UCenter), thus the
GCS displayed ground speed and the circling problem disappeard.

I am very pleased for the quick respond! You turned our day to a
success.

Regards,
  Balazs



2015-05-29 18:46 időpontban Gautier Hattenberger ezt írta:

> After looking at the nav code, your problem really seems to be the
> missing VELNED message in your GPS module conf.
>
> G.
>
>
> Le 29/05/2015 16:11, Balázs GÁTI a écrit :
>> Hi,
>>
>> Paparazzi version: v5.4.0_stable
>>
>> The two log files can be downloaded:
>>
>> ftp.edufly.net Port 21
>> user: [hidden email]
>> p: NW7100log
>>
>> Time of interest
>> 13_03_52: t= 315s
>> 13_47_23: t= 320s
>>
>>
>> Balazs
>>
>>
>>
>> 2015-05-29 15:43 időpontban Gautier Hattenberger ezt írta:
>>> Can you post the log files somewhere ? And give exact paparzzi
>>> version (./paparazzi_version)
>>>
>>> Gautier
>>>
>>> Le 29/05/2015 15:22, Balázs GÁTI a écrit :
>>>> Hi,
>>>>
>>>> I just experienced an interesting behavior of a Lisa/M v2
>>>> (Transition Robotics) with Paparazzi v5.4.0.
>>>>
>>>> I had finished tuning my fixed wing in AUTO1 and switched to
>>>> AUTO2. The plane started to intercept altitude and to align himself
>>>> to the Standby circle. After a half (more or less) stabilized circle
>>>> she terminated the circle and started to roll to the opposite
>>>> direction. I switched back to AUTO1 and landed. Post flight analysis
>>>> of log showed no anomaly. No warning, no lost RC or GPS, no GPWS
>>>> warning, no change in PPRZMode, motor running.
>>>>
>>>> I increased the height of waypoints and started a new flight. The
>>>> result was the same. After a half circle in Standby mode, she
>>>> started an opposite spiral. I left her to do it and experienced a
>>>> path similar to one and a half spiral in strong in wind by constant
>>>> roll. Post flight analysis showed a rapid change in Desired.Roll,
>>>> even the carrot and the green circle of the Standby mode still
>>>> behaved as I expected.
>>>>
>>>> The rapid change in Desired.Roll happened more or less in the same
>>>> time as Attitude.Psi changed sign. This is only expectation, because
>>>> this sign error would ruin all Paparazzi flight in the world. Even
>>>> my simulations didn't predict this behavior. Unfortunately the
>>>> telemetry messages are not too dense to confirm this expectation. Do
>>>> you have any advice what should I do to identify the reasons? (NEO7P
>>>> works already fine.)
>>>>
>>>>
>>>> Regards,
>>>>  Balazs
>>>>
>>>> _______________________________________________
>>>> Paparazzi-devel mailing list
>>>> [hidden email]
>>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>>
>>>
>>> _______________________________________________
>>> Paparazzi-devel mailing list
>>> [hidden email]
>>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>>
>> ---
>> Balazs GATI, PhD
>> associate professor
>> Department of Aeronautics, Naval Architecture and Railway
>> Vehicles
>> Budapest University of Technology and Economics
>>
>> Address:   Budapest
>> Stoczek u 6. J. ép. 423
>> 1111
>> Tel:       +(36)-1-463-1960 [1]
>> Fax:       +(36)-1-463-3080 [2]
>> Homepage: http://vrht.bme.hu/ [3]
>>
>> _______________________________________________
>> Paparazzi-devel mailing list
>> [hidden email]
>> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
>
>
> _______________________________________________
> Paparazzi-devel mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

---
Balazs GATI, PhD
associate professor
Department of Aeronautics, Naval Architecture and Railway
Vehicles
Budapest University of Technology and Economics

Address:   Budapest
Stoczek u 6. J. ép. 423
1111
Tel:       +(36)-1-463-1960 [1]
Fax:       +(36)-1-463-3080 [2]
Homepage: http://vrht.bme.hu/ [3]

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: eliminate ATT steady-state error

flixr
Administrator
In reply to this post by Eltonwu
IERROR_SCALE of 1024 seems a bit large indeed (too course quantization).
Sergey, could you please make a pull request with the reduced IERROR_SCALE?

On Wed, May 27, 2015 at 9:47 AM, Eltonwu <[hidden email]> wrote:
Hi all
ISCALE_ERROR reduce to 128 works perfect

Thank you very much



--
View this message in context: http://lists.paparazziuav.org/eliminate-ATT-steady-state-error-tp17019p17032.html
Sent from the paparazzi-devel mailing list archive at Nabble.com.

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel


_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
Reply | Threaded
Open this post in threaded view
|

Re: eliminate ATT steady-state error

flixr
Administrator
Hi,

I changed the IERROR_SCALE and adjusted the prescalers so that it should (tm) result in the igain having the same effect (except with less static offset due to smaller quantization error).

See https://github.com/paparazzi/paparazzi/pull/1205, would be nice to get some feedback on this.

Cheers, Felix

On Tue, Jun 9, 2015 at 12:01 AM, Felix Ruess <[hidden email]> wrote:
IERROR_SCALE of 1024 seems a bit large indeed (too course quantization).
Sergey, could you please make a pull request with the reduced IERROR_SCALE?

On Wed, May 27, 2015 at 9:47 AM, Eltonwu <[hidden email]> wrote:
Hi all
ISCALE_ERROR reduce to 128 works perfect

Thank you very much



--
View this message in context: http://lists.paparazziuav.org/eliminate-ATT-steady-state-error-tp17019p17032.html
Sent from the paparazzi-devel mailing list archive at Nabble.com.

_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel



_______________________________________________
Paparazzi-devel mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel