Page 444 - Sensing, Intelligence, Motion : How Robots and Humans Move in an Unstructured World

SUGGESTED COURSE PROJECTS 419

3. VLSI routing as a motion planning problem (designing a printed board is a
variation of this problem). Assume that after a VLSI chip has been already
designed, a need appears to add another few wires. With the traditional
tools, the whole design must be recalculated from scratch; this is expensive
and not always possible. Address the task as a motion planning problem.
The new wires cannot intersect those already on the chip, so the old wires
and other electronics on the chip present obstacles to new wires. Options
may include motion only in one plane or between planes, in which case
a wire may jump from plane to plane only in designated points.
4. Motion planning amongst moving obstacles. The topic can be formulated
for either a mobile robot or an arm manipulator. Investigate the constraints,
if any, to be imposed on the velocities and directions of motion of the
obstacles (relative to the robot) to make the problem tractable.
5. Mutual collision avoidance of two arm manipulators operating in shared
space.
6. Effect of arm kinematics on workspace accessibility in the presence of
obstacles. As we learned in Chapters 5 and 6, an interaction between
the arm and an obstacle creates shadows, which is effectively a part
of workspace that is not accessible to the arm. For the same size and
shape obstacle, the amount of workspace lost due to this effect depends
on the arm kinematics and relative dimensions of its links. Investigate
this phenomenon.
7. Algorithm for coordinating information from different sensors—for
example, in the context of sensitive skin sensors. As presented in
Chapter 8, each sensor of the skin looks exactly in front of it, catching the
reflection of its light from an obstacle in this direction. If, for example,
the obstacle’s surface is at an angle to this sensor, much of reflected light
will go in a different direction, possibly to another sensor. Therefore, the
fact of the object detection by both sensors may provide additional infor-
mation about the object’s shape, location, and distance from the robot
body. Investigate the coordinated use of sensor data for such and other
useful inferences about the obstacles.
8. Effect of robot dynamics on motion planning. This set of topics is good
for students with a strong background in control theory. The study of
effects of robot dynamics in Chapter 4 is done for a mobile robot that is
either a point, or a very symmetrical body.
• Investigate effects of dynamics for a mobile robot of a more com-
plex shape, e.g., a boxy four-wheel vehicle.
• Investigate effects of dynamics for an arm manipulator of a given kine-
matics.
9. Motion planning for highly redundant kinematic structures:
• Snakes
• Multi-finger wrists as multi-snake systems: power (whole-wrist) grasp-
ing, precision (two-point) grasping; pick-and-place operation, and so on