[Back]

S. Heuböck, G. Retscher:
"Investigation of the Performance of Low-cost HS-GPS Under Different Kinds of Forest Canopy";
Poster: IAG Scientific Assembly 2013, Potsdam, Germany (invited); 2013-09-01 - 2013-09-06; in: "IAG Scientific Assembly", (2013), 2 pages.

Current low-cost high sensitivity receivers (HS-GPS) technology can detect signals even under unfavourable visibility conditions. In agricultural applications for instance, GNSS receivers are already standard and often have to be used also in forests. This study investigates the practicability of low-cost HS-GPS under forest canopy.
A test area in the Wienerwald in the west of Vienna was selected which is a mixed forest with various variations of kind, age and form of trees in a relatively small area. In the test area ten survey points were established with conventional surveying. They are located in miscellaneous forests parts, i.e. in broad-leaved forest, coniferous forest and young forest. In our experiments long-term static measurements over a period of 24 hours were performed at each survey point. They were repeated three times in different seasons with different states of foliation. The first campaign was in summer with thick foliation, the second in October with thin foliation and the last one in March without leaves on the trees. In the study a u-blox LEA-6T receiver which also logs the raw data for post processing was used with its standard antenna. Each data set is processed and analysed in three different ways. First, a code single point solution using broadcast ephemeris and EGNOS corrections is calculated. Second, the data is post processed with precise ephemeris using single point positioning. Finally, relative positioning solutions using baselines to a virtual reference station in the Austrian EPOSA CORS network are calculated. These three different solutions are compared with the absolute coordinates of the survey points. In addition, two forest paths with a length of about 1.5 km and 500 m were measured to test the receiver in kinematic mode while driving with an off-road vehicle and while walking.
First results show that the HS-GPS availability is high in the forest areas under investigation. Kinematic measurements, however, showed devaitions of up to 20 m. For the static observation the code solution has a mean accuracy of less than five metres. Analysing the observations depending on the kind of canopy, it can be seen that the positioning results under leaves are worse than under coniferous canopy and the solutions in young forestry are the worst. The code solution using EGNOS is significantly better than the one without, even if EGNOS corrections are often not available during the measurement campaign.
The possibility of tracking the L1 signal also allows to carry out differential carrier phase measurements. It is an effective way to evaluate the performance of the receiver and improve the measurements under forest canopy. Due to the attenuation and delay by the foliation, however, the carrier phase information is heavily disturbed. The results can be divided in phase solutions where all ambiguities could be fixed and float solutions. The results show that less than one metre accuracy can be achieved in the case where the ambiguities could be fixed. On the other hand, three quarters of all solutions are float solutions and they have a large number of outliers. In forests with high canopy the percentage of succesful phase solutions is about 10, while under low canopy it is about 50 %. In young forests the deviations from the ground truth increase significantly compared to the code solutions, in broad-leafed and coniferous forests they decrease. First analyses even presume that the signal quality in summer with thick, soaked foliation is better than in autumn when the foliation is dry.
As the data analysis is still ongoing further test results will be presented in this contribution. It could already be proven how useful modern low-cost HS-GPS receivers are for applications like navigating or monitoring forest machines as well as for advanced operations such as cadastral boundary surveys or even monitoring of landslides. A typical application would be, for instance, determining the size of property land for agricultural European Union funding.