Data Set Overview : TRANSFORMED data files consist of time ordered records of Inertial Measurement Unit (IMU) quaternion elements and accelerations for the Backshell IMU; and quaternion elements and integrated change in velocity for the Rover IMU measured during the Entry, Descent, and Landing phases of the MER1 (Opportunity) and MER2 (Spirit) landers. For these files, the quantities have been (via onboard spacecraft processing) spatially transformed to the center of mass of the entry vehicle prior to backshell:lander separation, and following separation to the backshell's center of mass (for the Backshell IMU) and the center of mass of the rover (for the Rover IMU).  HIGHRATE data files consist of time ordered records of Inertial Measurement Unit (IMU) three-axis rotation rates and velocity changes for the Backshell IMU; and three-axis rotation rates and velocity changes for the Rover IMU measured during an approximately five minute subset of the approximately 30 minute Entry, Descent, and Landing phases of the MER1 (Opportunity) and MER2 (Spirit) missions. These HIGHRATE rotation rates and velocity changes have not been spatially transformed.   Parameters : The TRANSFORMED data files contain the following measurements: Spacecraft Clock Time: Units : seconds Rover IMU Quaternion Element 1: Units : N/A Rover IMU Quaternion Element 2: Units : N/A Rover IMU Quaternion Element 3: Units : N/A Rover IMU Quaternion Element 4: Units : N/A Backshell IMU Quaternion Element 1: Units : N/A Backshell IMU Quaternion Element 2: Units : N/A Backshell IMU Quaternion Element 3: Units : N/A Backehsll IMU Quaternion Element 4: Units : N/A Rover IMU Velocity Change X-Direction: Units : m/s Rover IMU Velocity Change Y-Direction: Units : m/s Rover IMU Velocity Change Z-Direction: Units : m/s Backshell IMU Acceleration X-Direction: Units : m/s**2 Backshell IMU Acceleration Y-Direction: Units : m/s**2 Backshell IMU Acceleration Z-Direction: Units : m/s**2  The HIGHRATE data files contain the following measurements: Spacecraft Clock Time: Units : seconds Rover IMU X-Direction Rotation Rate: Units : radians/s Rover IMU Y-Direction Rotation Rate: Units : radians/s Rover IMU Z-Direction Rotation Rate: Units : radians/s Rover IMU X-Direction Velocity Change: Units : m/s Rover IMU Y-Direction Velocity Change: Units : m/s Rover IMU Z-Direction Velocity Change: Units : m/s Backshell IMU X-Direction Rotation Rate: Units : radians/s Backshell IMU Y-Direction Rotation Rate: Units : radians/s Backshell IMU Z-Direction Rotation Rate: Units : radians/s Backshell IMU X-Direction Velocity Change: Units : m/s Backshell IMU Y-Direction Velocity Change: Units : m/s Backshell IMU Z-Direction Velocity Change: Units : m/s   Data : Each of the two Mars Exploration Rover (MER) spacecraft had two Inertial Measurement Units (IMU) on board. All four IMU were quality controlled Litton LN-200S units (http://www.ngnavsys.com/SiteMap/). The four units were selected from a larger population for improved stability. The IMU were on the spacecraft as engineering sensors, primarily to assist in Entry, Descent and Landing (EDL), but also to assist in surface mobility. Thus the data were processed and collected for engineering uses and have a number of unusual features.  The two IMU on each spacecraft were located in significantly different positions. One was inside the rover (the rover IMU: RIMU) and thus inside the lander and one was attached to the backshell (backshell IMU: BIMU). Note that neither IMU was particularly close to the entry capsule center of mass or spin axis. The data from the RIMU were collected throughout EDL (although there are gaps where they were not trasmitted back to Earth, see below). Just before bridle cut (right after the retro rockets begin firing and just before the bouncing begins), the BIMU is turned off since bridle cut 'discards' the backshell (and BIMU). See [CRISPETAL2003] for details on the spacecraft design and EDL timeline.  The very high rate data produced directly by the LN-200S units were summed in firmware and sampled at 8 Hz. In some cases the 8 Hz data were then stored directly and eventually played back. This is the HIGHRATE dataset. These data only cover a relatively short portion of the entire EDL process (since they were not of prime engineering interest). These data have been processed to physical units. This is due to the digital output of the IMU itself being essentially in physical units. The DN values from the IMU were converted to physical values on the spacecraft in the process of summing to the 8 Hz sampling. The ground depacketization then directly ended up with the physical units.  The TRANSFORMED datasets contain data that were used by the on-board algorithms in realtime during the EDL process. The on-board software converted the IMU output into physical units and then transformed them for use by the flight software. All of the transformations were based on the nominal (not measured) properties of the spacecraft (and various components). The accelerations were transformed to the center of mass (CM) and into the spacecraft coordinate system, changing the effective axes of the data. Before the deployment of the lander, this is the CM of the entry capsule. Afterwards, the BIMU measurements were transformed to the CM of the backshell and the rover IMU measurements to the CM of the lander. Secondly, the RIMU data in each axis were accumulated in a register to give accumulated delta-V since the start of the entry. The gyroscope data were used to compute a quaternion for the orientation of the spacecraft component relative to J2000. Like the acceleration data, this was done for the entire capsule before lander deployment for both IMU; and then for the backshell with the BIMU data, and lander for the RIMU data. The quaternions calculated onboard the spacecraft were positive normalized. Thus, the fourth element is redundant and was not transmitted back to Earth. During depacketization, the fourth element was recalculated: e4 : sqrt(1 - e1*e1 + e2*e2 + e3*e3) where e1..e4 are the four elements of the quaternion. This calculation produced 'complete' quaternions so they are easier to use. This calculation was performed for all the quaternions and results in the fourth element being set to unity when there are no data (since the other three elements are then zero).   Processing : The processing was done on the spacecraft for every one of the 8 Hz samples. But, due to limitations on the spacecraft memory, only some of the samples were stored and these samples were not co-added before being stored. The frequency of the saved data depended on the phase of the entry process (these range from 0.05 Hz to 4 Hz, see the actual datasets for the various sampling rates). In addition to being stored on board, some of the TRANSFORMED data were sent to MGS (Mars Global Surveyor) during the EDL process over the UHF link. This also continued for a short period after landing. Some of the data sent directly to MGS were also stored on board and some were not stored (also, some of the stored data were not sent over the UHF link). Due to the motions of the lander, the UHF link to MGS was somewhat intermittent and parts of the data sent were lost (there is also a regular data loss due to the realtime nature of the data transmission). This is the cause of the various gaps, holes and irregular spacing of the samples. The data returned during EDL and the stored data sent back later are complementary in many cases (in particular, for MER2 the two streams are both at 4 Hz but on alternating 8 Hz timesteps, resulting in an effective 8 Hz record--minus the lost data over the UHF link). Data from the two paths have been merged together in the archived datasets. This was done during the telemetry processing, before being delivered for calibration. The data from the two sources are identical when they overlap (since they are essentially the same data played back twice by different communication paths), and so this was only a removal of duplicate records (and sorting by time, but the order of the returned packets was not chronological in any case).  One other correction was made to the data during telemetry processing. Timesteps without any IMU data were removed (these were usually caused by events that caused a momentary loss of the UHF link). There were some timesteps for which only a subset of the IMU data (usually the quaternion element values) were transmitted. In these instances, the file record includes the transmitted values, but the non-transmitted values for these records have been set to a value of zero. Note that zero is a legal value for an IMU measurement. These instances of incomplete IMU data transmission are identified by the occurrence of all six accelerometer values (three RIMU and three BIMU) having the value of zero, or all six rotation rate values in the HIGHRATE files, or eight quaternion element values in the TRANSFORMED files, having the value of zero.  The accumulated delta-V values for the RIMU were reset to 0 by the flight software while those data were used during EDL (the cause is known and the timing of the resets is known due to the spacecraft events they are associated with). This is the cause of the sudden jump in the RIMU data where they go to zero in all three axes and then start accumulating once again.  The data presented are those transmitted from the spacecraft. Since processing was done onboard the spacecraft (since these data were used in real-time by the onboard software to provide control and critical event triggering during EDL), the 'raw' DN values were not stored on the spacecraft. Rather, the values used by the on-board software (acceleration, accumulated velocity changes, rotation rates) are the values that were transmitted and are available for this archive.

Confidence Level Overview :   Data Coverage and Quality : The complete 8 Hz data collected and utilized during EDL are not available within this archive. Data transmission dropouts between each rover and MGS have resulted in several short gaps within the data.