HUMAN FOLLOWING ROBOT USING KINECT SENSOR

Abstract: This study presents a three-wheeled human following robot with a

Kinect sensor as its only sensor for following a human target and avoiding
collision. It is the goal of the study to eliminate certain limitations for human
following robots such as limited operating area and requiring the user to wear
intrusive sensors. The robot follows a user by detecting the hip of the user and
acquiring the depth data of the hip for steering and maintaining distance.
Several tests are made to characterize the robot, these test are: straight path
test, varying curve radius test, and collision avoidance test. Though the robot
was able to follow the user in a straight path, it was noticed that there is an
average deviation of 4 to 8 cm. from the path of the user. In the varying curve
test, the safest turning radius is at least 1.5 meters, which enables the robot to
not collide with the vertex of the arc made by the user and not lose track of the
user. In this test, the robot also tends to undercut the user. For collision
avoidance tests conducted, the robot was successful to avoid obstacles but it
makes, at most, a 180 cm. of deviation from the path of the user. Although the
robot is able to follow a user in a straight path; the robot tends to undercut the
user during curved path. There is also a possibility that the robot will follow a
different user after losing the user in its field of view during the target
reacquisition process.

Development of a Controller Board with On-board Sensors for RotorBased Unmanned Aerial Vehicles

Abstract: Unmanned aerial vehicles (UAV) research works have grown

immensely over the past decades due to its maneuverability and vast possible
applications. These applications are always dependent on its controller board
and on-board sensors. This research aims to develop a controller board
equipped with various sensors such as an Inertial Measurement Unit (IMU) for
stability of the roll and pitch, a magnetometer for stability with regards to yaw,
a barometric pressure sensor and an ultrasonic sensor for height, and a GPS for
location mapping. Filters are used for IMU, barometric pressure sensor and as
well as the ultrasonic sensor for a well-defined reading. These sensors are
necessary to achieve a stable flight for a rotor-based unmanned aerial vehicle
for four, six and eight rotors. A remote controller is incorporated to the
controller board for future developments when it is used in actual flight. This
can be used as a switch to fly the UAV autonomously, or maneuver the UAV

from one place to another. Such controller board must be able to control at
most eight motors using PWM signals sent to the ESC. The controller board
would allow further studies to incorporate control systems to the board. The
traces of the controller board was designed using the software EAGLE and its
components were placed specifically to control a UAV. It uses an ATMEGA2560
microcontroller and is programmed by Arduino using C language. The different
sensors will be controlled by the microcontroller using different techniques such
as Serial and I2C. It is then implemented to an octocopter to ensure that the
controller board could control 8 simultaneous motors.

Integrated Electronic Toll Collection System Using RFID and GSM

Technology

Abstract: The electronic toll collection system in the Philippines at the time of
this research uses microwave technology. It is dubbed as e-pass, it allows
motorists to pass through the toll gates by simply slowing down until the
system detects the signal. However, microwave detectors are expensive.
Hence, the full deployment is not practicable. As the number of vehicles that
use the tollgates increases, the need to fully implement the e-pass system is
desirable. RFID offers a cheaper alternative.
The Integrated Electronic Toll Collection System using RFID and GSM
Technology (ETC RFID) developed in this study has four main functions: (1) to
detect the RFID tag of a vehicle that enters the tollgate, (2) to detect the RFID
tag of a vehicle that exits the tollgate, (3) to send the data read by the reader
to the database server and (4) to notify the balance of the customers RFID tag
through SMS. The graphical user interface that shows the computation of the
toll fee is displayed using Visual Basic6.0.
The ETC RFID system is implemented in a pseudo toll gate. The system
comprises the reader, the tag, the GSM module, a local PC containing the GUI
for the logs and a central server that contains the database. The reader is
placed in the pseudo toll gate, while the tag is installed inside the vehicle.
Several tests were done to determine the optimum angle of the reader with
respect to the tag. The angles are 300, 450, 600, and 900. It is observed that
as long as the speed of the vehicle is 50 kph or less, the position of the reader
does not affect its detecting capability. A 100% success rate is achieved.
However, when the speed is increased, only the 450 angle position enabled the
reader to detect the tag installed in the vehicle that is running at a maximum
speed of 80 kph.

The GSM capability test shows in all trials hundred percent success rate in
sending the notification whenever the customers account is below the set
limit.

Automatic Flood Detection Using the Video of Static Cameras

Abstract: Regional heavy rainfall is usually caused by extreme weather

conditions. Consequently, heavy rainfall often results in the flooding of rivers
and the neighboring low-lying areas, which is responsible for a large number of
casualties and considerable property loss. The existing weather forecast
systems focus primarily on the analysis and forecast of large-scale areas and
do not provide precise instant automatic monitoring and alert feedback for
small areas and sections. In this paper, we propose a method to automatically
monitor the flood level of a specific area based on video surveillance systems
using image processing methods and computer vision techniques to provide
instant feedback on flooding events. OpenCV, which specializes in real-time
computer vision, is used for motion tracking and image processing. The system
can track objects and analyze if its flood based on a set of features. The
proposed method can better meet the practical needs of disaster prevention
than large-area forecasting. It also has several other advantages, such as
flexibility in location deployment and it does not require specialized equipment.
The results offer prompt information and reports for appropriate disaster
warning and response in small localized areas.

Developing Robot Motion Models and Gesture Recognition System for a

Human-Robot Interaction System to assist in Play Therapy

Abstract: Studies show that children with low-functioning Autism Spectrum

Disorder (ASD) have problems with their motor skills and social development.
Play therapy is a technique used to aid the development of children with
special needs through play, as play helps children develop cognitive, social and
perceptual skills. This study aims to assist therapists in the process through a
follow-the-leader type game with an interactive robot system. This paper
tackles the results of two main experiments: the data extraction of natural
human motion as basis of robot motion model and the evaluation of Hidden
Markov Model (HMM) gesture recognition algorithm to enable human-robot
interaction. Natural movement is defined as a robot trying to achieve humanlike motions as close as possible. This process can be done by collecting a total
of fifteen sample video of human actions, tracking specific joints and

manipulating the robots actuator such that the velocity, acceleration, and
torque are controlled based on the action provided. The primary goal is to
identify whether the robot can reach arm accelerations of 5m/s2 to 20m/s2 and
-5m/s2 to -20m/s2. The arm should also reach elbow angles ranging 35o to 50o
during the transition of the robots arm toward its body and 86o to 116o during
the transition away from the body. The values are taken from the conducted
experiments to produce the desired output with natural human movement. A
direct attempt to reproduce the same movement patterns from the data
collected from the test subjects produced a lot of jitter as the robot tries to
move as fast as possible such as reaching the peak acceleration of 20m/s2 and
-20m/s2. Aside from producing natural movement, the system also needs to
detect and recognize gestures that are given by the child. Gesture recognition
requires machine learning algorithm such as HMM to be able to train gestures
efficiently. To evaluate the HMM gesture recognition algorithm, a total of 65
gesture data per gesture were collected to use for training and testing. The
HMM algorithm yielded low accuracy rate on determining the Lateral Raise
gesture wherein only one out of five Lateral Raise were correctly classified
while Hand Wave has four out of five. The result could be influenced by the
small training set or the complexity of the gestures which made it hard to
attain more than 50% accuracy rating.

Data Management System of an FLC-Based Temperature Monitoring for

Air Conditioning System

Abstract: An integrating home automation system is now becoming a trend in

modern house concepts. These systems may include the function of remotely
switching devices/appliances. Different technologies are being utilized to
achieve this function for houses termed as intelligent home. Using the
related studies which are gathered and further analysed by the proponents,
they have come up with the idea of creating a simple model for a remote
switching application system. The proponents choose to develop a model using
a telephone system. Telephone line is chosen to serve as the channel for the
transmission of instruction signals. The system will be composed of a remote
section and a local section. DTMF signalling, which is the standard dialling
technique for telephone system, will be used as a means of transmitting of the
signals to the local section. The generated signals from the control section will
be decoded using a DTMF decoder. The brain of the system is an Arduino
microcontroller. This will take care of the processing of the main functions of
the system.

A Model Case of Intelligent Data Management and Decision Supporting

System of Monitoring Green Leafy Vegetable

Abstract: SMART farming is an important undertaking as increasing demand in

agricultural industries in the Philippines is notable. Smart farming in the
country could empower local farmers with the decision tools and technologies
needed for their livelihood. However, this technology is not yet available in the
country. For sustainable agriculture in midst of climate change, monitoring of
plant health and detecting its diseases is critical. The need to effectively grow a
plant and increase its productivity should be given emphasis. In addressing
these concerns, monitoring and grading the plant at its growth and at the time
of its harvest is necessary. The purpose of this paper is to provide a model of
monitoring the plants and suggest solutions to the farmers for healthy yield
and productivity. In this work, the proponent proposed a model to be used in
monitoring the identified vegetables during farming. The model for grading
vegetables is divided into three complex phases: a) design and development of
a plant disease detection system using machine vision, b) design and
development of an artificial intelligent system in vegetable grading, and c)
notification to the farmer of plant grade and status. Two vegetables have been
selected. The images are to be interpreted numerically and it will be classified
in linguistics. This paper will demonstrate effective algorithms for vegetable
monitoring and grading. Practical implementation of fuzzy logic will be done
using MATLAB. For this purpose, Sugeno-style of fuzzy inference system will be
used.