Agriculture Reference
In-Depth Information
red color difference and luminance values. It was determined that the red color dif-
ference values were much more effective at detecting the apples than the luminance
values. Bulanon et al. (2009) demonstrated improved citrus fruit detection through
multiperspective viewing of a fixed boundary region of interest, achieving approxi-
mately 90% detection rates in orange canopies.
Numerous other sensors are commonly used in robotic harvesting systems, such
as ultrasonic range, laser range, capacitive proximity, and light-emitting diode range.
It is not likely that a single sensor will solve the complete sensing problem; rather,
several sensors will need to be integrated together to form a complete sensory system.
7.4.2.3 Robotic Manipulation and Control
The manipulator is defined as a mechanical system, usually composed of a series of
actuated links that function like a human arm capable of moving within 1-D, 2-D, or
3-D space. In robotic fruit harvesting, the tool end of the manipulator is fitted with a
fruit gripper, whereas the base is typically mounted on a mobile platform that posi-
tions it in the tree canopy. The manipulator's task is to move the gripper into posi-
tion to pick the fruit and then place the fruit into a collection bin. The manipulator
is composed of joints and links, similar to the human arm, with each joint having 1
degree of freedom (DOF). In general, a 3-DOF robot can provide maneuverability to
any point in 3-D space without regard to orientation, whereas a 6-DOF manipulator
can move to any point within its work space volume with complete position and ori-
entation capabilities. However, a 6-DOF system is generally limited to a single pose
and as such may not be able to avoid an obstacle in its workspace. If 3-D position
and orientational tool frame accuracy is required, a redundant manipulator may be
required. A redundant manipulator must have 1 more DOF than the degrees of posi-
tioning accuracy required. Therefore, if
X
,
Y
,
Z
Cartesian position and orientational
degrees of pitch, roll, and yaw are required, a minimal redundant manipulator would
have 7 DOF with additional degrees being optional. However, numerous research
and development efforts have been implemented that used less than 7 DOF, often
suffering for lack of maneuverability to avoid obstacles. Although additional degrees
of freedom can improve maneuverability and tool frame dexterity, they also increase
manipulator cost and control complexity. The manipulator's geometric configura-
tion or architecture, forward and inverse kinematic algorithms, and the manipulator
dynamic equations of motion form the key design characteristics of a robotic manip-
ulator. Unlike assembly lines in a factory, harvesting fruit trees (apples, oranges, etc.)
is highly unstructured, and the robot must have the workspace reach and end-effector
dexterity necessary to reach fruit within a complex environment cluttered with limb
and leaf canopy obstacles. There are several geometric configurations used in indus-
trial applications, shown in Figure 7.22, which have been applied to fruit harvesting:
Cartesian, cylindrical, spherical, articulated, and redundant.
The earliest laboratory prototype was an apple harvester (Parrish and Goksel,
1997) consisting of a simple arm with a pan-and-tilt mechanism and a touch sensor
in place of an end-effector, which made contact with modeled fruit. The first field
prototype for harvesting apples was developed in France (Grand D'Esnon, 1985).
The mechanical system consisted of a telescopic arm that moved up and down in a
vertical framework. The arm was mounted on a barrel that could rotate horizontally.
