Model of Floor Cleaner Robot
The model of FCR in SolidWork:
Figure 1: Model of FCR in SolidWorks
The model of FCR in real:
Figure 2-5: Model of FCR in real
Results and discussion
The FCR is able to successfully navigate the way to complete
the floor cleaning task. It also maintains the orbit that is programmed when
doing the task. Moreover, it also success in avoid some obstacles in the
testing environment and can be auto stop when the finishing the task. However,
compass measurement errors sometime take the FCR more time than expected to
turn in right direction. Besides that, the mechanical design and PID algorithm
errors also affect to FCR, make it does not move in a straight line.
In the first test case, a room with no obstacle, the FCR can
be work in stable. It can be cleaning almost the area of the room, except some
area near the wall. When increasing the obstacle in the room, the FCR also
completes the task. But it can be seen that, in more and more complicated
environment with many obstacles, the area that not be cleaned will be increase.
So it is the big challenge that must be resolved in the future to make the FCR
become more stable and can adapt any environment.
The video of testing FCR prototype can be found at the following link:
Future works
After constructing and experiment the model of Floor Cleaner
Robot, it can be seen that the FCR can adapt in some ease testing environment.
And to make this model become more stable and adapt any working environment,
there are many things to do in the future. This section discusses of the robot
have not yet been implemented, but should be to complete the full functionality
of FCR.
Structure
- Self-preservation
An important feature that needs to be implemented for FCR
success is that of knowing when it is in danger. This danger can come either
from external sources, such as a pet or stairs, or internal sources, such as
battery drain. Although relatively straight forward in nature, each of these
problems can cause complications hindering overall performance.
If a curious puppy or a vindictive cat decides to investigate
the robot, how will it handle itself? Although it is very low to the ground and
stable on its own right, the pet could easily push the robot off its course.
The robot can use the infrared sensor to determine when the pet or human come
to and have the solution such as turn off system to protect human and pet or
try to avoid by create the new path working. Of course, a determined animal
will not be deterred by these actions, so the robot would then just aim to be
still until the animal got bored. Ideally, nay owner would first acclimate the
pet to the robot’s presence before letting it operate without supervision.
As for the presence of stairs in the environment, adding
another short-range distance sensor to the front of the robot looking downwards
would detect changes in elevation well before the robot was in danger of
passing over it. Assuming that the robot is initialized on a flat surface, it
can then calibrate its own sensor to determine the default value. Then any
deviation can be recognized as an indicator of physical danger and avoidance
routine instated
A more complicated matter is how to monitor and conserve
battery levels. There are chips on the market that provide an indicator of the
voltage output from a battery pack. One in particular is MAX832 chip, produced
by Maxim, which can switch between two separate battery packs when necessary.
This chip can also recharge the batteries when is senses the presence of an AC
power source. Recognizing when battery levels are low is relatively straight
forward using this off-the-shelf microchip. Getting the robot find an available
AC source on its own, however, is a whole other problem.
- Reduce size and weight of FCR
The problems of FCR are very big and heavy. FCR is difficult
to clean all area of room. It cannot move under obstacles such as bed, table,
and cabinet. Another hand, reducing the weight of FCR is necessary.
Solution:
- Using small battery instead of
big battery.
- Change the design or form of FCR
Algorithm
- Returning to home base
Having the robot know where it came from and able to return
to that same location is very difficult problem. There are several paths to
finding a solution, but determining which path depends on several philosophical
factors. For instance, the robot could “remember” where it came from, either by
building a map, recording the set of moves that it took so get where it was, or
by leaving breadcrumbs. What these solutions have in common is that they all
require some form of memory and for meaningful information to be retained for
significant duration. If the decision is made not to rely on complex memory
structures, an organic approach could be implemented by installing a beacon of
sorts that draws the robot back to its starting point, also the location of its
battery charger.
In keeping with the smile to nature that is behavior-based
AI, one must first decide if simple creatures use memory or beacons to find
their home after a long day exploring. Whether animals use landmarks in the
environment, remember the movements of their body, or leave signs for
themselves along their path varies from species to species. The complexity of
these techniques also varies, as does the reliability. For instance, human are
known to navigate by landmarks, but then, we also have very large brains that
can remember the details of a wide variety of objects and how to differentiate
them from other, similar objects. Honeybees can communicate the locations of
food sources several miles away by way of a complicated “dance” that conveys information
about which direction to fly and when to turn. And the humble ant is able to
lead others to food and to find its way back to its colony by depositing
pheromones along its trail
Although the method used by the ant to modify its
environment in a way only it personally could is simplest of the three, it is
not acceptable for a robot to alter its surroundings in a physical, and
therefore sense-able, way. A more practical method might be to install a beacon
of sorts at the home location whose attraction supersedes that of sunlight as
the batteries run down. Grey Walter took this approach when he was designing
his MachinaSpecualtrix. As the
batteries ran down however, the photocells became less sensitive and more
brilliant sources became attractive. By design, the robots’ hutches were lit by
a bright bare light bulb. Thus the hutches became the most attractive area in
the environment, and the robots appeared to know to go home for sustenance when
their energy levels were depleted. FCR can use an RF beacon to be guided it
back to its home location based on signal strength.
