Targeting and image acquisition of Martian surface features with TGO/CaSSIS

CaSSIS is a high-resolution visual telescope onboard the ExoMars Trace Gas Orbiter. The mission started the primary science phase in April 2018. The relatively small single image footprint (typically 40 km x 9.5 km) when compared to the total surface area of Mars demands that images should be targeted and target selection is key for the science return. This Jo urn l P repro of paper describes the science planning concept set around the target selection, and the process followed in order to generate the CaSSIS commands. The tools used are described as well as all the iterations and teams involved. Finally, special cases and the handling of contingencies are discussed. The procedures may serve as a guideline for future highresolution instruments on missions to planetary objects.


Introduction
ExoMars [1] is a Mars exploration program consisting of two missions, an orbiter launched in 2016 and a rover mission planned to be launched in 2028. The 2016 mission, also known as the Trace Gas

Orbiter (TGO), was launched in March 2016 and, after orbit insertion in
October 2016 and a period of aerobraking, entered in its primary science phase in April 2018. The payload of TGO is designed to perform observations of the atmosphere of Mars using two high resolution spectrometers, ACS [2] and NOMAD [3], along with imaging of J o u r n a l P r e -p r o o f potential trace gas source regions at moderately high resolution.
Additional objectives include the study of transient, dynamic phenomena on the surface [1] and the mapping of hydrogen in the sub-surface down to a depth of approximately 1 metre [4].
The Colour and Stereo Surface Imaging System (CaSSIS) is the imager on board TGO and has been designed to characterise sites which have been identified as potential sources of trace gases, to investigate dynamic surface processes, and characterise potential future landing sites [5]. Typically, images are around 9.5

km x 40 km in size at a pixel scale of about 4.5 m/px from the nominal orbit altitude and in 3 or 4 colours.
Stereo images can also be acquired using a rotation mechanism. The data rates from the spacecraft limit the number of image acquisitions to between 1 and 4 per orbit (12- CaSSIS can image the surface. The constraints and requirements implied by the above leads to a complex planning process for CaSSIS that is needed to optimize the data acquisition and target the most interesting places on Mars under good illumination conditions. This paper describes the operational planning procedure and is intended to support future imaging instrument operations by indicating how the scientific return has been optimized using multiple tools. This information also helps users of the science data to understand the dataset.
In the next three sections, we shall recall the important aspects of the instrument and the spacecraft that lead to the planning requirements including the targeting methodology. In the following section, we shall outline the planning methodology used by ESA that was required to be supported. In section 6 we shall describe our planning tools and their capabilities. In sections 7, 8, we shall describe the deliverables and the verification of the execution. We conclude with a summary emphasizing the benefits of the system used but also identifying some of the aspects that could be improved upon for future missions.
J o u r n a l P r e -p r o o f

The CaSSIS instrument
A full description of CaSSIS is given in [5]. Here a summary of the instrument is given focusing on elements of CaSSIS that have a direct impact on the planning process of CaSSIS images.
CaSSIS is a stereo multi-colour push-frame imager based upon an F/6.5 telescope with a focal length of 880 mm. The detector is a 10 μm pixel pitch CMOS hybrid detector system originally developed for the SIMBIOSYS imager for BepiColombo [6]. CaSSIS has an angular scale of 11.36 μrad/px which, when combined with the nominal orbit altitude (roughly circular 400 km above the surface) leads to a spatial scale on the surface of Mars of approximately 4.5 m/px. In order to obtain multi-band information of the acquired sites, filters were placed on top of the CMOS detector as shown in Figure 1. The images are obtained by aligning the line that is orthogonal to all filters with the spacecraft velocity vector over the surface of the planet (i.e. the spacecraft ground track, as it can be seen in Figure 2). To achieve this, CaSSIS has a motor that can be commanded to a defined position prior J o u r n a l P r e -p r o o f  Figure 2).

J o u r n a l P r e -p r o o f
The images acquired can cover the full swath (left to right in Figure 1) or only part thereof. A trade-off is made between swath width and the number of colours acquired. In a push-frame system, images are acquired at a rapid frequency that matches the speed at which the spacecraft flies over the surface. The framelets generated by this rapid imaging must subsequently undergo a mosaicking process to produce the final image.

Target Suggestion
The target suggestion follows a similar process that was developed and refined by the HiRISE team [9] and ported to CaSSIS.  (Chojnacki et al., 2020). In this context, the discussion provided here should be useful for future proposers to optimize the suitability of

J o u r n a l P r e -p r o o f
Note that all constraints added at this stage will strongly influence the possibilities for the target to be imaged and therefore the time it will take to actually acquire the image.  [11]. The bulk of the planning for the instruments is performed during the MTP and STP parts of the cycle as can be seen in    J o u r n a l P r e -p r o o f

The suggestion interface, CaST
CaST is the CaSSIS target proposal tool. Based on the HiRISE [12] tool HiWish [9], it allows users to propose targets for CaSSIS to image through a modern web interface (Figure 4) J o u r n a l P r e -p r o o f   [13], and provides a similar service to the HiPlan tool used by the HiRISE operations team [12].
Within PLAN-C, two custom data layers are created: a target database layer and an image planning layer.     A phase angle correction is needed only when the phase angle is less than 20°, so the global average Mars photometric function of Vincendon et al. (2013) [16]

The Ground Reference Model (GRM) and The Electronic Ground Support Equipment (EGSE)
In order to improve the quality and quantity of data the CaSSIS

J o u r n a l P r e -p r o o f
This provides a rapid visual assessment of the current status of CaSSIS and includes displays of the major voltages and currents within the instrument, the temperatures of key components, command counters, the mechanism position, and internal memory usage (see Figure ). Command acceptance and failure can also be tracked.

Evolution of the CaSSIS Planning
At the start of the science phase, CaSSIS already had a robust planning system but many aspects of it were manual. Finally, once the target has been acquired, the "retirement" of the target request within the target database is challenging [19]. In particular, the non Sun-synchronous orbit, dynamic phenomena (e.g. dust storms on Mars), and instrument anomalies can influence the success or failure of J o u r n a l P r e -p r o o f an observation and this can only be judged by the requestor. As a result, science team interaction is a vital part of the process.
The full end to end process is viewed in the schematics of Figure that includes all the tools and interactions described in the paper. Starting with CaTS and CaTL selecting the targets to be planned from the CaST database that was populated by the CaTL beforehand. And the CTF and ITL products being sent to ESA where they are made into attitude and instrument commands that populate the Mission Timeline (MTL) that contains all the commands that are ultimately executed on the spacecraft and produce the planned images at exact locations. These are returned to Earth in a process described in [19] that includes a feedback mechanism that retires acquired targets from the database. MGS