Instrument Overview
  ===================
    The Rock Abrasion Tool (RAT) is an integral part of the Athena
    Science payload; it was produced to act as the payload's geologic
    hammer and more. The RAT exposes fresh surfaces of martian rocks
    to other instruments on the payload. The RAT also brushes dust and
    debris from an excavated hole or an unaltered rock target in
    addition to its rock hammer equivalency role. To accomplish these
    tasks, the RAT is a sophisticated 3-axis precision-controlled
    device that will allow it to act also as a rock physical
    properties science instrument. The returned RAT data will be
    inverted and compared to a library of Earth rocks.
 
    The RAT is 128 mm long and is contained within a circle 85mm in
    diameter and has a mass of 687 grams.
 
    Information in this instrument description is taken from the Rock
    Abrasion Tool Mars Exploration Rover mission paper
    [GOREVANETAL2003]. See this paper for more details.
 
 
  Scientific Objectives
  =====================
    The chief scientific objectives of the RAT are:
 
    1) to remove the outer, exposed surface of rocks,
 
    2) to provide data such as motor currents so torque can be
       directly calculated and empirically correlated to the density and
       hardness of the rock and any coatings that have formed on the
       rock, and
 
    3) to collect information on the mineralogic and magnetic properties
    of the abraded rock by collecting magnetic dust on four magnets
    mounted on the lower part of the RAT housing.
 
 
  Calibration
  ===========
    It is necessary to calibrate the RAT prior to each grinding
    operation in order to compensate for certain time-variant or
    temperature-dependent parameters. Calibration begins with the
    initialization for the RAT Z-axis. This is required to correct for
    undetectable migration of the Z-axis that may be caused by
    vibration experienced during Rover traverses. To initialize the
    Z-axis, the motor is 'homed' using its hard stop, thus
    establishing an absolute reference frame for subsequent position
    moves. Next, the Rotate motor 'no load current' is
    calibrated. This is necessary to account for changes in drive
    train friction, lubricant viscosity and dust buildup over time and
    temperature range. To calibrate the Rotate motor 'no load'
    current, both the Rotate and Revolve motors are energized at
    operational levels for a preset period of time. Sensor values are
    averaged over this period and updated parameters are stored for
    subsequent use.  Likewise, the Rotate motor 'seek-scan voltage' is
    calibrated in the same manner. It is preferred, but not required,
    that the RAT be calibrated in free space, before being placed
    against a rock surface.  This minimizes the chance of encountering
    obstructions during calibration.
 
 
  Operational Considerations
  ==========================
    The RAT is required to be able to grind 5 mm below the surface of
    a rock after it has been properly positioned against an
    appropriate rock surface. An appropriate rock surface is
    considered a surface with local deviations no greater than +/-5
    mm within a 50 mm radius from the point where the RAT Z-axis
    intersects the surface plane. It is the responsibility of the
    control of the Instrument Deployment Device (IDD) to position the
    RAT against an appropriate rock surface such that the RAT's Z-axis
    is within +/-15% from the rock's surface normal.  The IDD must
    preload the RAT against the rock surface with a force no less than
    10N and no greater than 100N. Once the RAT has been positioned on
    the rock, the grinding process is handled completely by the RAT with
    no additional manipulation required by the IDD or Rover.
 
    Several important operational constraints were levied upon the RAT
    grinding operation. First, the pose of the (IDD) must not be
    disturbed during the grinding operation. Second, the grinding
    operation must not expend more than an allotted amount of energy.
    Third, the grinding operation should take no longer than a preset
    time limit. The RAT control approach was driven primarily by these
    operational requirements and constraints.  An emphasis was placed on
    flexibility and modularity, so that an operator can easily tune the
    grinding operation by changing key parameters.
 
 
  Electronics
  ===========
    From an actuator/sensor perspective, the RAT is an electrical
    interface similar to other devices (IDD, mobility) on the
    Rover. The RAT has three DC brush motors, and each motor has an
    incremental quadrature encoder for position feedback. The RAT also
    has redundant discrete switches to sense contact with the
    rock. Finally, the RAT has a temperature sensor and motor heaters
    to monitor the mechanism's temperature and ensure that the RAT is
    within operational temperature limits prior to beginning
    operations.
 
    The RAT relies on the Rover control electronics for low-level
    servo control and sensor processing. The Rover motor controllers
    allow the motors to be commanded in one of three modes: servo to
    position, servo to velocity and open-loop voltage mode. The
    low-level servo sample rate is on the order of 1 kHz. Motor
    current sensing is also provided by the Rover control
    electronics. Current limits are enforced at multiple levels
    (controller and application software) to prevent damage to the
    actuators, grinding wheels and IDD.
 
 
  Location
  ========
    The RAT is mounted on the end of the IDD.
 
 
  Measured Parameters
  ===================
    The Rock Abrasion Tool's main function on Mars is to remove the
    outer, exposed surface of rocks. In order to accomplish this task,
    the RAT is designed to cut approximately 5 mm into the rock and
    expose the fresh, unweathered surface. During the grinding
    operations, the RAT records data such as motor currents, motor
    positions, temperature, and instrument state. Torque can be
    directly calculated from motor current data and empirically
    correlated to the density and hardness of the rock. Combining
    these parameters with a close-up view of the fresh surface from
    the Microscopic Imager (MI), the scientist should be able to
    assess the compositional and the depositional histories of the
    rock.