Author Contributions
The idea for the power self-calibration algorithm as well as the clock drift correction was done by J.S.; The coauthors V.S., N.S.-N., M.A. and U.H. improved the paper through their comments and corrections about the layout, content and the obtained results. Conceptualization, methodology, software, validation, visualization, formal analysis, investigation and writing-original draft preparation was done by J.S.; Supervision, project administration was done by N.S.-N. and U.H.; Funding acquisition was done by M.A.
Figure 1.
Example of the phase locked loop (PLL).
Figure 1.
Example of the phase locked loop (PLL).
Figure 2.
Frequency difference between the received carrier and the internal phase locked loop (PLL) in parts per million (PPM). The curve is obtained by reading the carrier integrator value of the DW1000 chips. The logarithmic increase of the curve is due to the warm-up of the crystal oscillator.
Figure 2.
Frequency difference between the received carrier and the internal phase locked loop (PLL) in parts per million (PPM). The curve is obtained by reading the carrier integrator value of the DW1000 chips. The logarithmic increase of the curve is due to the warm-up of the crystal oscillator.
Figure 3.
Filtered frequency difference between the received carrier and the internal phase locked loop (PLL) in parts per million (PPM). The colors represent different measurement obtained successively. It can be seen that the curves are deterministic.
Figure 3.
Filtered frequency difference between the received carrier and the internal phase locked loop (PLL) in parts per million (PPM). The colors represent different measurement obtained successively. It can be seen that the curves are deterministic.
Figure 4.
Frequency difference during the warm-up phase of the DW1000 crystal oscillator in parts per million (PPM) with respect to the time in minutes [
15]. The figure is used with permission [
15].
Figure 4.
Frequency difference during the warm-up phase of the DW1000 crystal oscillator in parts per million (PPM) with respect to the time in minutes [
15]. The figure is used with permission [
15].
Figure 5.
Signal power in dBm of the received blink message. The curve shows the filtered results of the received signal power over time, measured by the DW1000 chip. After 4600 measurements the transmitting signal power was reduced.
Figure 5.
Signal power in dBm of the received blink message. The curve shows the filtered results of the received signal power over time, measured by the DW1000 chip. After 4600 measurements the transmitting signal power was reduced.
Figure 6.
Filtered frequency differences between the received carrier and the internal phase locked loop (PLL) in parts per million (PPM) after changing the transmitting signal power after about 2340 s.
Figure 6.
Filtered frequency differences between the received carrier and the internal phase locked loop (PLL) in parts per million (PPM) after changing the transmitting signal power after about 2340 s.
Figure 7.
The measurement setup consist of two transceivers (EVK100) with the identification number 1 and 2.
Figure 7.
The measurement setup consist of two transceivers (EVK100) with the identification number 1 and 2.
Figure 8.
Schematic for the presented clock drift correction based on three transmitting messages. See text for explanation.
Figure 8.
Schematic for the presented clock drift correction based on three transmitting messages. See text for explanation.
Figure 9.
Filtered received signal power measurements of the three messages: P1, P2 and P3.
Figure 9.
Filtered received signal power measurements of the three messages: P1, P2 and P3.
Figure 10.
The error in meters caused by the clock drift. The curve is changing, due to the clock warm-up.
Figure 10.
The error in meters caused by the clock drift. The curve is changing, due to the clock warm-up.
Figure 11.
The error in meters after the clock drift correction. .
Figure 11.
The error in meters after the clock drift correction. .
Figure 12.
The effect of the received signal power on the distance measurement [
15], used with permission. The red line represents the correct distance measurements. The blue line shows the range bias caused by different signal powers [
15].
Figure 12.
The effect of the received signal power on the distance measurement [
15], used with permission. The red line represents the correct distance measurements. The blue line shows the range bias caused by different signal powers [
15].
Figure 13.
Measured received signal power at the DW1000 chip, with respect to the correct signal power. The blue line represent the reference line, every value of the x-axis has the same value on the y-axis. The other lines are the results of the estimated received signal power with different settings. The estimated received signal power equates the correct signal power for low signal strength [
14].
Figure 13.
Measured received signal power at the DW1000 chip, with respect to the correct signal power. The blue line represent the reference line, every value of the x-axis has the same value on the y-axis. The other lines are the results of the estimated received signal power with different settings. The estimated received signal power equates the correct signal power for low signal strength [
14].
Figure 14.
Filtered received signal power measurements of the three messages P1, P2 and P3. The measurements have been obtained with a cable connection between the transceivers. The transmitted signal power of the second signal P2 is reduced with a step size of 3 dB, while the signal power for P1 and P3 remains constant.
Figure 14.
Filtered received signal power measurements of the three messages P1, P2 and P3. The measurements have been obtained with a cable connection between the transceivers. The transmitted signal power of the second signal P2 is reduced with a step size of 3 dB, while the signal power for P1 and P3 remains constant.
Figure 15.
Filtered timestamp error changes due to the received signal power. The measurements have been obtained with a cable connection between the transceivers. The error is changing systematically with decreasing signal power.
Figure 15.
Filtered timestamp error changes due to the received signal power. The measurements have been obtained with a cable connection between the transceivers. The error is changing systematically with decreasing signal power.
Figure 16.
Filtered received signal power measurements of the three messages P1, P2 and P3. The measurements have been obtained with a wireless connection between the transceivers. The transmitted signal power of the second signal P2 is reduced with a step size of 0.5 dB, while the signal power for P1 and P3 remains constant.
Figure 16.
Filtered received signal power measurements of the three messages P1, P2 and P3. The measurements have been obtained with a wireless connection between the transceivers. The transmitted signal power of the second signal P2 is reduced with a step size of 0.5 dB, while the signal power for P1 and P3 remains constant.
Figure 17.
Filtered timestamp error changes due to the received signal power. The measurements have been obtained with a wireless connection between the transceivers. The error is changing with decreasing signal power.
Figure 17.
Filtered timestamp error changes due to the received signal power. The measurements have been obtained with a wireless connection between the transceivers. The error is changing with decreasing signal power.
Figure 18.
Interference between the received messages P1, P2 and P3. The transmitted signal power of the message P1 and P3 are the same. Due to the short update time are the signals interfering.
Figure 18.
Interference between the received messages P1, P2 and P3. The transmitted signal power of the message P1 and P3 are the same. Due to the short update time are the signals interfering.
Figure 19.
Estimated correction curve between the measured received signal power and the ideal signal power curve.
Figure 19.
Estimated correction curve between the measured received signal power and the ideal signal power curve.
Figure 20.
Correction value of the error caused by the signal power as a function of the received signal power.
Figure 20.
Correction value of the error caused by the signal power as a function of the received signal power.
Figure 21.
Power correction curve for the received signal power.
Figure 21.
Power correction curve for the received signal power.
Figure 22.
Correction value of the error caused by the signal power as a function of the received signal power for a different station with six restarts.
Figure 22.
Correction value of the error caused by the signal power as a function of the received signal power for a different station with six restarts.
Figure 23.
Schematic for the signal power correction for two-way ranging.
Figure 23.
Schematic for the signal power correction for two-way ranging.
Figure 24.
Two-way ranging with clock drift correction between reference station and tag.
Figure 24.
Two-way ranging with clock drift correction between reference station and tag.
Figure 25.
Difference between the measured distances obtained by two-way ranging and the ground truth distances in meter. The constant offset caused by the hardware delay is not compensated but is represented by the red line. The near most constant errors for different distances shows the power correction is correct.
Figure 25.
Difference between the measured distances obtained by two-way ranging and the ground truth distances in meter. The constant offset caused by the hardware delay is not compensated but is represented by the red line. The near most constant errors for different distances shows the power correction is correct.
Table 1.
Test settings.
Parameter | Value |
---|
Center Frequency | 3993.6 MHz |
Bandwidth | 499.2 MHz |
Pulse repetition frequency | 64 MHz |
Preamble length | 128 |
Data rate | 6.81 Mbps |
Table 2.
Notations used.
Notations | Definition |
---|
| Timestamp |
| Clock drift with respect to the timestamps n and m |
| Timestamp error due to signal power |
Z | Hardware delay and signal power correction offset |