GPS Field Experiment for Balloon-based Operation Vehicle

GPS Field Experiment for Balloon-based Operation Vehicle P.J. Buist, S. Verhagen, Delft University of Technology T. Hashimoto, S. Sakai, N. Bando, JAXA p.j.buist@tudelft.nl 1 Objective of Paper This paper will describe a GPS-based precise relative positioning and attitude determination experiment on the Balloon-based Operation Vehicle. In section 2, background information on the Balloon-based Operation Vehicle is provided and section 3 will provide more information about the GPS experiment. Preliminary results of the first phase of the experiment are presented and our plans for the second phase are discussed. 2 Introduction On Earth, micro-gravity experiments can be conducted in drop towers or dropped from high altitudes. To eliminate the effect of the air drag, the inside of the drop tower has to be vacuum or, if the experiment is dropped from some vehicle, the drag has to be compensated. ISAS (Institute of Space and Astronautical Science, presently a part of JAXA) has been developing balloon-dropped experimental systems since 1978. However, at that time, the balloon technology was not mature enough to fly over 4 km altitude, and the system did not have a propulsion system which could cancel the air-drag. Nowadays, the balloon technology has matured, and JAXA obtained the world-record for high altitude balloon technologies [1]. Consequently, since 24, JAXA has continued to develop a Balloon-based Operation Vehicle (BOV), and performed experiments in 26 (BOV1) and 27 (BOV2) [2]. Fig. 1 shows the BOV during ascent and after it was returned to the launch site in the 27 experiment. Main goal of the project is the development of a micro-gravity test system, but the BOV is also a good environment to test other technologies. For example in 29 a test of an air-breathing engine in supersonic flight is planned. For this experiment a wing-type BOV will be applied and it will pull up from freefall after which it is expected to reach a velocity of more than 1.5 Mach. JAXA has been using GPS for coarse positioning of the BOV and the balloon, but next would like, in cooperation with Delft University of Technology (DUT), to test methods for GPS-based precise relative positioning and attitude determination. Fig. 1 BOV2 during launch and after return to the launch site 2.1 Typical Flight Sequence of Balloon Fig. 2 shows a typical flight profile of a high altitude balloon from JAXA. Normally, a so called gondola is hanging under the balloon, at which experiments can be conducted. In about 2 hours the balloon climbs to its maximum altitude of 35-4 km. At this altitude it will stay for about 4 hours, until by telecommand the exhaust valve of the balloon is opened and the balloon descents into the ocean. The maximum velocity the balloon reaches is about 4 km/h during descent. Altitude (km) 4 35 3 25 2 15 1 5 Altitude Velocity 6: 7: 8: 9: 1: 11: 12: 13: Local Time (hr) 2 18 16 14 12 1 8 6 4 2 Velocity (km/h) Fig. 2 Typical Flight profile of Gondola from GPS data

2.2 Typical Flight Sequence of BOV (26 & 27) The BOV is brought to the desired altitude using a high-altitude balloon. The balloon lifts the payload with a mass of about 7 kg to an altitude of 4 km in about 4 hours. At the start of the experiment, a telecommand is send to the balloon to disconnect the BOV and start the microgravity experiment during the free-fall of the BOV. During the microgravity experiment, the microgravity condition is maintained inside the BOV by controlling the relative distance from the BOV to the microgravity capsule within the BOV with a cold gas propulsion system. This distance is measured using laser range sensors. The experiment descents around 2 to 3 km in about 3 seconds after which the microgravity condition is terminated. Next the parachute is deployed and the BOV lands in the ocean. Within 2 hours, the BOV is retrieved by ship/ helicopters. Fig. 3 shows position and velocity of the BOV from the release from the balloon until it splashes into the ocean. The maximum velocity it reaches is about 15 km/h. Altitude (km) 5 45 4 35 3 25 2 15 1 5 Altitude Velocity - 5 5 1 15 2 25 3 35 Time from Separation (s) Fig. 3 Flight Profile BOV2 from GPS data 2 18 16 14 12 1 2.3 BOV Future Plans In 28, the JAXA balloon center has moved from Sanriku, Iwate to Taiki-cho, Hokkaido. Now the 3rd flight of the standard BOV, as shown on fig. 1, is scheduled for 29 (TBD). The 4th flight of the BOV will test an air-breathing engine in supersonic flight (29 TBD). This flight utilizes a wing-type BOV. During the flight, there will be a pull up from freefall, after which the vehicle reaches a velocity of more than 1.5 Mach. For this flight a new attitude determination package (ADP) is under development [3]. This package contains a sun sensor, a GPS sensor, a magnetic sensor, a gyro, and an inclinometer (instrument for measuring angles of inclination of an object with respect to gravity). The GPS receiver in this package is solely used for positioning. 8 6 4 2 Velocity (km/h) Fig. 4 Wing-type Balloon-based Operational Vehicle (BOV) 3 GPS Experiment 3.1 GPS-based Precise Relative Positioning and Attitude Determination High-end GPS receivers make use of two types of observations: pseudo range and carrier phase. Low cost receivers, as for example used in car navigation systems, only make use of pseudo range observations. The pseudo range observations typically have an accuracy of decimeters, whereas carrier phase observations have accuracies up to millimeter level. The drawback of carrier phase observations is that they are inherently ambiguous by an unknown integer number of cycles, and therefore we will have to apply an ambiguity resolution method before it can be used for positioning or attitude determination. Therefore a reliable ambiguity estimation strategy is the key factor for both precise relative navigation and attitude determination. The LAMBDA method, developed at DUT, is considered the standard for a wide range of GNSS ambiguity resolution problems [4]. For a baseline constrained application, as for example GPS-based attitude determination, we can make use of the knowledge that the length of the baseline is known and constant [5]. For GPS-based precise relative navigation, two basic approaches are available: 1 kinematic approach 2 Extended Kalman Filter (EKF) approach In the EKF approach, a dynamics model is used in combination with observations, whereas the kinematic approach is solely based on GPS observations. The main advantage of the kinematic approach over EKF is that information about dynamics of the system is not applied, which gives more flexibility and furthermore could improve the scientific interest of the observations made by the mission. Therefore a functional kinematic approach is still of high importance for future space missions, but also for terrestrial applications as for example airborne gravimetry and formation flying with UAVs and aircraft. Precise relative positioning of two moving platforms usually requires dual-frequency phase data, whereas due to the baseline length constraints - single-frequency phase

data may suffice for the precise determination of platform attitudes [6][7][8]. Recently good results were achieved using single epoch ambiguity resolution between a reference station and moving vehicles as a vessel and aircraft [8]. Additional field experiments, in cooperation with the Japanese space organization JAXA, are planned to validate these strategies in the BOV environment. 3.2 GPS Experiment on BOV For the GPS experiment we would like to determine the relative positions between the elements and the attitude of the elements of the experiment. There are number of options available to place the GPS receivers. It would be possible to have a single or multi- antenna GPS receiver on a reference point at the ground, on the balloon s gondola and/or the BOV. This would make it possible to do precise relative positioning between a known point (the reference station at the ground) and one or two other elements, or even between the other elements (gondola and BOV). The multi antenna GPS systems on gondola and/or BOV would make it possible to do attitude determination of these elements. An impression of the experiment is shown in fig. 5. GNSS Satellite BOV <>Gondola Balloon Gondola BOV will make a large attitude maneuver from freefall to the horizontal flight path and will reach very high velocities by accelerating from free fall, both effects will have an impact on the data from the GPS receiver. As the relative motion between the elements of the experiment is unpredictable, the kinematic approach is the most suitable processing strategy for relative positioning. Furthermore the antennas are placed under the balloon, which will affect the GPS signals. 3.4 Description of the first phase of the experiment In the first phase of the experiment we will investigate the Balloon Environment. The first phase was started in the summer of 28. On the 4 th of September two GPS receivers were flown on the gondola of the high altitude balloon. Goal of the experiment was to collect data from a single baseline for offline analysis. This data will provide us with more information about blocking and reflection of the GPS signals on the gondola. The data from the single baseline can be used to determine the pointing direction of the gondola (an application referred to as the GPS Compass [7]). The data is collected on the single baseline with 2 dual frequency GPS (&GLONASS) receivers (see fig. 6), coined com1 and com2 after the communication (or com) port used to connect them to the controller. The receivers are controlled by, and the data stored on an ARM-Linux [9] over a RS232 connection. The receivers and the controller are stored in a water proof container as shown in fig. 6. The location of the two GPS antennas on the gondola is shown in fig. 7. The baseline length is 1.95 meter. The ADP package under development for the wing-type BOV was also flown on the same flight to test it in a representative environment [3]. Ref. station <>BOV Ref. station <>Gondola Ref. station Fig. 5 Impression of second phase of BOV experiment 3.3 Challenges of Experiment The BOV is a challenging environment for precise GPS-based relative positioning as it is actively controlled during free fall to avoid the collision between the outer shell (the capsule body itself) and the inner shell (micro-gravity experimental module) by a cold gas propulsion system. JAXA has the world-record of high altitude balloon technologies and also the high altitude of the experiment is a challenge as the BOV payload can reach an altitude higher than 4km. For the GPS experiment this is a very interesting altitude as not many experiments have been performed at this height, which is higher than the altitude reachable by an aircraft but below Low Earth Orbits for spacecraft. Furthermore the BOV Fig. 6 Container with the ARM-Linux and two GNSS receivers

GPS antennas Fig. 7 GPS antennas on the Gondola 3.5 Preliminary results of the first phase The number of locked channels for both receivers is shown in fig. 8. Also the altitude from the GPS receiver prepared by JAXA s balloon group (hereafter coined as the onboard receiver) is included in this figure. The figure shows that between 4 and 12 satellites are locked on both receivers, even before take off and after landing in the ocean. The number of locked satellites is different for both receivers and this should be due to differences in the local environment of the antennas. The balloon itself contains some metal objects, for example the exhaust valve, which can cause blocking and reflections of the GPS signals. Fig. 9 shows the altitude of the gondola estimated from the data collected with the two GPS receivers plus the altitude estimated by the onboard GPS receiver. This figure proves that we were able to collect raw data during the whole flight, including the phase before take of and after landing, and at the highest altitude of about 4 kilometer. In this balloon flight, we had to lead the balloon eastward to the ocean for safe recovery. Therefore, the ascent of the balloon was stopped at around 12 km, an altitude where the wind blew strongly eastward. At this altitude, around at 7:am, the exhaust valve was opened to reduce the lift and stop the ascent. Then, around 7:3am, the ballast was dropped and the balloon re-ascended. Fig. 9 Altitude based on post- processing of raw data Next we will estimate the relative position vector between the two antennas. For this we will solve the carrier phase ambiguities using the standard LAMBDA method [4] as described in [8]. We will use single epoch data for ambiguity fixing. After we find this baseline vector we will express the results in a heading (or azimuth) angle which is defined from the North direction and an elevation angle which is defined from the horizontal plane. In these figures, we used the data after the exhaust valve was opened and until landing in the ocean. With (x,y,z), the baseline vector in East, North, Up frame, the definition of the two angels is as follows: y Azimuth( = heading) = atan z Elevation = atan x z + y 2 2 The results for heading are shown in fig. 1 and for elevation in fig. 11. Similar as presented in [8], we use a validation method based on the ratio test on the fixed ambiguities. When the ratio test is valid, we accept the fixed solution; otherwise we will keep the float solution. From the heading estimation it is clear that the gondola rotates around its vertical axis until it reaches equilibrium. From their on it rotates in the opposite direction until it reaches a new equilibrium. From the elevation angle we can observe that, as expected, the gondola remains mostly horizontal during the flight, and thus this result confirms that the fixed solution is correct. From the figure it is also clear that the precision of the float solution is lower than the fixed solution. Fig. 8 Number of locked GPS satellites

test environment for our kinematic approach for relative positioning in combination with attitude determination. The objective is to demonstrate the capability for relative positioning and attitude determination for the purpose of guidance and control. Acknowledgment The MicroNed-MISAT framework is kindly thanked for their support. Fig. 1 Heading based on post- processing of raw data Fig. 11 Elevation based on post- processing of raw data 3.6 Description of the second phase of the experiment For the second phase of the experiment we would like to collect dual baseline data from a number of vehicles. The preferred option for the experiment is 3 antennas / 2 baselines on the BOV, and 2 antennas / 1 baseline at the gondola. Furthermore a number of antennas at a reference station could be used to represent another platform in our network. The goal of the second phase of the experiment is to collect data for offline relative navigation & attitude determination, though real-time emulation, between a number of vehicles which form a network. This second part of the experiment is planned for 29. An impression of the experiment was shown in fig. 5. 4 Conclusions Innovative algorithms for relative positioning and attitude determination are under development at Delft University of Technology. Up-until-now, the algorithms were tested with data collected on a vessel, an aircraft and at static reference stations. The Balloon-based Operation Vehicle from JAXA is a good environment to test in a more dynamic environment. Especially the flight of the wing-type BOV will have unpredictable motion due to cold gas control during free fall, a large attitude maneuver to horizontal flight and acceleration during the testing of the air breathing engine. This is anticipated to be a good References [1] T.Yamagami, Y. Saito, Y. Matsuzakaa, M. Namikia, M. Toriumia, R. Yokotaa, H. Hirosawaa and K. Matsushima Development of the highest altitude balloon, Advances in Space Research, Volume 33, Issue 1, pp.1653-1659, 24 [2] T. Hashimoto, S. Sawai, S. Sakai, N. Bando, H. Kobayashi, K. Fujita, Y. Inatomi, T. Ishikawa, T. Yoshimitsu, Y. Saito, Progress of Balloon-based Micro-gravity Experiment System, 26th International Symposium on Space Technology and Science (ISTS), Hamamatsu, Japan, 1-8 June, 28 [3] T. Hashimoto, S. Sawai, S. Sakai, N. Bando, H. Kobayashi, Y. Inatomi, T. Ishikawa, K. Fujita, T. Yoshimitsu, Y. Saito, H. Fuke, S. Shimizu, P. J. Buist, Development of High Altitude Balloon-based Micro Gravity Experimental System (Progress Report for FY 28) (in Japanese), 25-26 September 28, Balloon Symposium 28, Sagamihara, Japan [4] P.J.G. Teunissen, A New Method for Fast Carrier Phase Ambiguity Estimation, IEEE Position, Location and Navigation Symposium PLANS 94, Las Vegas, US, April 1994 [5] P.J.G. Teunissen, The LAMBDA Method for the GNSS Compass, Artificial Satellites, Volume 41, Number 3, 26, pp. 89-13. [6] C. Park, P.J.G. Teunissen, A new carrier phase ambiguity estimation for GNSS attitude determination systems, Proceedings of International GPS/GNSS Symposium, Tokyo, pp. 283-29, 23 [7] P.J. Buist, The Baseline Constrained LAMBDA Method for Single Epoch, Single Frequency Attitude Determination Applications, ION-GNSS 27, Fort Worth, Texas, US, September, 27 [8] P.J. Buist, GNSS Kinematic Relative Positioning for Spacecraft: Data Analysis of a Dynamic Testbed, 26th International Symposium on Space Technology and Science (ISTS), Hamamatsu, Japan, 1-8 June, 28 [9] http://armadillo.atmark-techno.com/armadillo