Limits...
Improving the accuracy of estimates of animal path and travel distance using GPS drift ‐ corrected dead reckoning

View Article: PubMed Central - PubMed

ABSTRACT

Route taken and distance travelled are important parameters for studies of animal locomotion. They are often measured using a collar equipped with GPS. Collar weight restrictions limit battery size, which leads to a compromise between collar operating life and GPS fix rate. In studies that rely on linear interpolation between intermittent GPS fixes, path tortuosity will often lead to inaccurate path and distance travelled estimates. Here, we investigate whether GPS‐corrected dead reckoning can improve the accuracy of localization and distance travelled estimates while maximizing collar operating life. Custom‐built tracking collars were deployed on nine freely exercising domestic dogs to collect high fix rate GPS data. Simulations were carried out to measure the extent to which combining accelerometer‐based speed and magnetometer heading estimates (dead reckoning) with low fix rate GPS drift correction could improve the accuracy of path and distance travelled estimates. In our study, median 2‐dimensional root‐mean‐squared (2D‐RMS) position error was between 158 and 463 m (median path length 16.43 km) and distance travelled was underestimated by between 30% and 64% when a GPS position fix was taken every 5 min. Dead reckoning with GPS drift correction (1 GPS fix every 5 min) reduced 2D‐RMS position error to between 15 and 38 m and distance travelled to between an underestimation of 2% and an overestimation of 5%. Achieving this accuracy from GPS alone would require approximately 12 fixes every minute and result in a battery life of approximately 11 days; dead reckoning reduces the number of fixes required, enabling a collar life of approximately 10 months. Our results are generally applicable to GPS‐based tracking studies of quadrupedal animals and could be applied to studies of energetics, behavioral ecology, and locomotion. This low‐cost approach overcomes the limitation of low fix rate GPS and enables the long‐term deployment of lightweight GPS collars.

No MeSH data available.


An example of gold standard and down‐sampled GPS paths, dead reckoned (DR), and two drift‐corrected dead reckoned (DCDR) paths from a domestic dog. This example is taken from domestic dog 5, trial 2. The gold standard GPS path has a sampling frequency of 3600 fixes per hour (1 Hz) and a true distance travelled (dt) of 2.43 km. The down‐sampled GPS path shows the problem of using low sample rate GPS data for distance estimation. It was created by down‐sampling the gold standard GPS path to 12 fixes per hour. The apparent distance travelled (da) calculated from this path is 1.99 km. The DR path drifts away from the gold standard GPS path as errors accumulate with time. This drift is reduced by applying a linear correction to the DR path between GPS position fixes (DCDR paths). The drift correction points are marked using magenta circles. Using points taken at 2 and 12 fixes per hour makes little difference to the apparent distances travelled (2.43 and 2.57 km, respectively), and hence, the proportional accuracy values are similar (0.99 and 1.05, respectively). Using drift correction with the higher GPS position fix rate, however, improves the ability of the dead reckoned path to follow the gold standard GPS path. For this example, DCDR with drift correction at 2 fixes per hour has a 2D‐RMS position error of 54.3 m; with 12 fixes per hour position error is reduced to only 12.9 m.
© Copyright Policy - creativeCommonsBy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC5016644&req=5

ece32359-fig-0005: An example of gold standard and down‐sampled GPS paths, dead reckoned (DR), and two drift‐corrected dead reckoned (DCDR) paths from a domestic dog. This example is taken from domestic dog 5, trial 2. The gold standard GPS path has a sampling frequency of 3600 fixes per hour (1 Hz) and a true distance travelled (dt) of 2.43 km. The down‐sampled GPS path shows the problem of using low sample rate GPS data for distance estimation. It was created by down‐sampling the gold standard GPS path to 12 fixes per hour. The apparent distance travelled (da) calculated from this path is 1.99 km. The DR path drifts away from the gold standard GPS path as errors accumulate with time. This drift is reduced by applying a linear correction to the DR path between GPS position fixes (DCDR paths). The drift correction points are marked using magenta circles. Using points taken at 2 and 12 fixes per hour makes little difference to the apparent distances travelled (2.43 and 2.57 km, respectively), and hence, the proportional accuracy values are similar (0.99 and 1.05, respectively). Using drift correction with the higher GPS position fix rate, however, improves the ability of the dead reckoned path to follow the gold standard GPS path. For this example, DCDR with drift correction at 2 fixes per hour has a 2D‐RMS position error of 54.3 m; with 12 fixes per hour position error is reduced to only 12.9 m.

Mentions: Figure 5 shows an example of a ground truth and a down‐sampled GPS path, dead reckoned, and two drift‐corrected dead reckoned paths (with drift correction at 2 and 12 fixes per hour). This example is taken from dog 5, trial 2. The ground truth GPS path has a sampling frequency of 3600 fixes per hour (1 Hz). The true distance travelled (dt) calculated from this path is 2.43 km. The down‐sampled GPS path is shown to illustrate the problem of using low fix frequency GPS data for distance estimation (Fig. 5). It was created by down‐sampling the ground truth GPS path to a rate of 12 fixes per hour. The apparent distance travelled (da) calculated from this path is 1.99 km, an underestimation of 0.44 km (a proportional accuracy of 0.82). The dead reckoned path drifts away from the ground truth GPS path as errors in the speed and heading estimates accumulate with time (Fig. 5). This drift is reduced by applying a linear correction between GPS drift correction fixes taken at regular intervals. The results of drift correction at intervals of 2 and 12 fixes per hour are shown in Figure 5. Drift correction points are marked using magenta circles. The apparent distance travelled for the drift‐corrected dead reckoned paths with drift correction at 2 and 12 fixes per hour is 2.43 and 2.57 km, respectively (proportional accuracy of 0.99 and 1.05, respectively). Using the higher drift correction sample rate of 12 fixes per hour improves the reconstruction to more accurately match the ground truth GPS path (Fig. 5) (2D‐RMS position error of 54.33 and 12.94 m at 2 and 12 fixes per hour).


Improving the accuracy of estimates of animal path and travel distance using GPS drift ‐ corrected dead reckoning
An example of gold standard and down‐sampled GPS paths, dead reckoned (DR), and two drift‐corrected dead reckoned (DCDR) paths from a domestic dog. This example is taken from domestic dog 5, trial 2. The gold standard GPS path has a sampling frequency of 3600 fixes per hour (1 Hz) and a true distance travelled (dt) of 2.43 km. The down‐sampled GPS path shows the problem of using low sample rate GPS data for distance estimation. It was created by down‐sampling the gold standard GPS path to 12 fixes per hour. The apparent distance travelled (da) calculated from this path is 1.99 km. The DR path drifts away from the gold standard GPS path as errors accumulate with time. This drift is reduced by applying a linear correction to the DR path between GPS position fixes (DCDR paths). The drift correction points are marked using magenta circles. Using points taken at 2 and 12 fixes per hour makes little difference to the apparent distances travelled (2.43 and 2.57 km, respectively), and hence, the proportional accuracy values are similar (0.99 and 1.05, respectively). Using drift correction with the higher GPS position fix rate, however, improves the ability of the dead reckoned path to follow the gold standard GPS path. For this example, DCDR with drift correction at 2 fixes per hour has a 2D‐RMS position error of 54.3 m; with 12 fixes per hour position error is reduced to only 12.9 m.
© Copyright Policy - creativeCommonsBy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC5016644&req=5

ece32359-fig-0005: An example of gold standard and down‐sampled GPS paths, dead reckoned (DR), and two drift‐corrected dead reckoned (DCDR) paths from a domestic dog. This example is taken from domestic dog 5, trial 2. The gold standard GPS path has a sampling frequency of 3600 fixes per hour (1 Hz) and a true distance travelled (dt) of 2.43 km. The down‐sampled GPS path shows the problem of using low sample rate GPS data for distance estimation. It was created by down‐sampling the gold standard GPS path to 12 fixes per hour. The apparent distance travelled (da) calculated from this path is 1.99 km. The DR path drifts away from the gold standard GPS path as errors accumulate with time. This drift is reduced by applying a linear correction to the DR path between GPS position fixes (DCDR paths). The drift correction points are marked using magenta circles. Using points taken at 2 and 12 fixes per hour makes little difference to the apparent distances travelled (2.43 and 2.57 km, respectively), and hence, the proportional accuracy values are similar (0.99 and 1.05, respectively). Using drift correction with the higher GPS position fix rate, however, improves the ability of the dead reckoned path to follow the gold standard GPS path. For this example, DCDR with drift correction at 2 fixes per hour has a 2D‐RMS position error of 54.3 m; with 12 fixes per hour position error is reduced to only 12.9 m.
Mentions: Figure 5 shows an example of a ground truth and a down‐sampled GPS path, dead reckoned, and two drift‐corrected dead reckoned paths (with drift correction at 2 and 12 fixes per hour). This example is taken from dog 5, trial 2. The ground truth GPS path has a sampling frequency of 3600 fixes per hour (1 Hz). The true distance travelled (dt) calculated from this path is 2.43 km. The down‐sampled GPS path is shown to illustrate the problem of using low fix frequency GPS data for distance estimation (Fig. 5). It was created by down‐sampling the ground truth GPS path to a rate of 12 fixes per hour. The apparent distance travelled (da) calculated from this path is 1.99 km, an underestimation of 0.44 km (a proportional accuracy of 0.82). The dead reckoned path drifts away from the ground truth GPS path as errors in the speed and heading estimates accumulate with time (Fig. 5). This drift is reduced by applying a linear correction between GPS drift correction fixes taken at regular intervals. The results of drift correction at intervals of 2 and 12 fixes per hour are shown in Figure 5. Drift correction points are marked using magenta circles. The apparent distance travelled for the drift‐corrected dead reckoned paths with drift correction at 2 and 12 fixes per hour is 2.43 and 2.57 km, respectively (proportional accuracy of 0.99 and 1.05, respectively). Using the higher drift correction sample rate of 12 fixes per hour improves the reconstruction to more accurately match the ground truth GPS path (Fig. 5) (2D‐RMS position error of 54.33 and 12.94 m at 2 and 12 fixes per hour).

View Article: PubMed Central - PubMed

ABSTRACT

Route taken and distance travelled are important parameters for studies of animal locomotion. They are often measured using a collar equipped with GPS. Collar weight restrictions limit battery size, which leads to a compromise between collar operating life and GPS fix rate. In studies that rely on linear interpolation between intermittent GPS fixes, path tortuosity will often lead to inaccurate path and distance travelled estimates. Here, we investigate whether GPS‐corrected dead reckoning can improve the accuracy of localization and distance travelled estimates while maximizing collar operating life. Custom‐built tracking collars were deployed on nine freely exercising domestic dogs to collect high fix rate GPS data. Simulations were carried out to measure the extent to which combining accelerometer‐based speed and magnetometer heading estimates (dead reckoning) with low fix rate GPS drift correction could improve the accuracy of path and distance travelled estimates. In our study, median 2‐dimensional root‐mean‐squared (2D‐RMS) position error was between 158 and 463 m (median path length 16.43 km) and distance travelled was underestimated by between 30% and 64% when a GPS position fix was taken every 5 min. Dead reckoning with GPS drift correction (1 GPS fix every 5 min) reduced 2D‐RMS position error to between 15 and 38 m and distance travelled to between an underestimation of 2% and an overestimation of 5%. Achieving this accuracy from GPS alone would require approximately 12 fixes every minute and result in a battery life of approximately 11 days; dead reckoning reduces the number of fixes required, enabling a collar life of approximately 10 months. Our results are generally applicable to GPS‐based tracking studies of quadrupedal animals and could be applied to studies of energetics, behavioral ecology, and locomotion. This low‐cost approach overcomes the limitation of low fix rate GPS and enables the long‐term deployment of lightweight GPS collars.

No MeSH data available.