PENSAR JUNTS, TREBALLAR PLEGATS

In order to well perform the operations of surveillance, face detection, recognition as well as subject tracking and navigation, the robot was designed in a certain way to meet the requirements of the tasks at hand.

The requirements were to construct the robot tall enough to be able to capture subject faces from a front perspective rather than from below as would be the case with a ‘short’ or a small robot. Also, the robot should have sufficient speed to be able to navigate to a subject in a relatively short time. These two requirements necessarily resulted in a third obligation, which is to make the robot stable so that it doesn’t fall when it moves, turns and stops during its navigation operation. The following subsections, present in some detail the robot’s body, hardware and camera system’s design.

3.1.1 Body

The robot’s body was constructed in three steps:

1. Robot’s base 2. Robot’s upper body 3. Improvements

The robot’s base was borrowed from a large radio controlled vehicle sized 1:6 of the original. This approach simplified and speed up the robot’s building process due to the fact that the base body, driving motors and wheels all come pre- assembled and tested for correct operation. Figure 3.1 below shows the base of the robot.

Figure 3.1: The robot’s base 30.5 cm

The robot’s body was built from angle slot aluminum having dimensions of 18x18 mm and a cross section of 0.5mm. This angle slot metal was chosen for its light weight as well as its relative strength which made it suitable for the robot’s operation. The metal slot was cut to form the robot’s body. Figure 3.2 shows the design and dimensions of the robot’s body.

Figure 3.2: The design and dimensions of the robot’s body

116.5 - 130 cm

135.3 - 148.8 cm Holding Clips

for height adjustment

After building the robot, several improvements were applied to enhance the robot’s operation thus giving it variable height as well as better stability during motion and navigation operations.

(a)

(b)

Figure 3.3: The improvements that were applied to the robot’s base and body:

As can be seen in Figure 3.2, the robot’s height can be manually adjusted, making it possible to extend or contract it to suit the operation environment. Other improvements are shown in Figure 3.3 (a) where the front bumper is being attached to a bent metal strip to improve its crash handling, so it minimizes the crash force without breaking, as well as a couple of Styrofoam blocks to protect the ultrasonic sensors during a crash. Figure 3.3 (b) shows the applied improvements to the robot’s suspension and wheels. The suspension was enforced to eliminate tilting during navigation, as well as the addition of a metal strip between the two wheels to limit their separation due to the robot’s body and components weight. Figure 3.3 (b) also shows the improvements that were applied to the wheels, where a hard plastic cover was put on the wheels to eliminate vibrations due to the ridges which affected the cameras’ picture quality.

3.1.2 Hardware

The robot’s hardware consists of sensors, controllers and actuators. It is responsible for maintaining correct operation during the robot’s navigation towards a subject allowing it to avoid obstacles and move in the desired path:

 Sensors: Four ultrasonic sensors were used to detect obstacles in the robot’s path.

 Controllers: Three Arduino controllers and one DC motor controller were used to receive signals from the sensors and send commands to the actuators.

 Actuators: Two DC motors were used to drive the robot in the desired direction.

The hardware system also included a small breadboard that was used to connect the ultrasonic sensors to Arduino board as well as host the authentication key (DIP switch) which was used in the authentication stage. In that stage, An Arduino controller would read the switch’s value and send it to the computer which will in turn compare it with the password entered by the subject during the verification procedure. Figure 3.4 shows the ultrasonic sensors, Figure 3.5 (a) shows the

controllers, while Figure 3.5 (b) shows the robot’s operational block diagram. Finally, Figure 3.6 shows the actuators (DC motors) that are used to drive the robot.

Figure 3.4: The ultrasonic sensors installed on the robot Upper sensors

(a)

(b)

Figure 3.5: (a) The controllers used on the robot, (b) The block diagram Arduinos

12v Battery DC Motor Controller

Figure 3.6: The actuators (dc motors) used on the robot

3.1.2.1 Power and Performance

The robot is equipped with a 12 volts battery, as can be seen in Figure 3.5. This battery is the power source for the dc motors that drive the robot as well as the Arduino board that drives the servo motors in the camera system mentioned in section 3.1.3. The rest of the robot hardware (microcontrollers and cameras) are powered by the laptop’s USB ports which supplies 5 volts.

The 12 volts battery provided sufficient power to drive the robot at a speed of around 1.9 meter per second. This speed was tested more than once with the robot fully loaded with all its operational equipment.

3.1.3 Software

The program used to control the robot operation was written in C++ using Microsoft Visual Studio 2010 under Microsoft Windows 7 – 64 bit version. The program listing is mentioned in Appendix A.

The written program made use of the open source computer vision software library known as OpenCV [75]. The open source software was used without being modified.

3.1.4 Camera System

The camera system used by the robot was designed to monitor the area surrounding the robot so as to detect any intruder that may enter into the robot’s operation area.

The system consists of three Genius F100 wide angle cameras and one full- HD normal angle camera. The three wide angle cameras each has a 120 degrees field of view, thus cooperatively covering 360 degrees, while the full-HD camera is used to capture the detected subject’s face image to be later used for face recognition.

The two rear wide angle cameras are stationary, while the front wide angle camera is stationary during the monitoring operation, but is fixed on top of a servo motor, thus has the ability to rotate right and left during the navigation operation. The full-HD camera is affixed to two servo motors, one rotating horizontally and the other rotating vertically, thus enabling moving the camera left, right, up and down to track the intruding subject. All the cameras used were connected to the main controller (laptop) through USB. A USB hub was utilized to facilitate connecting all the cameras to a single USB port on the laptop. Figure 3.7 shows the camera system used on the robot.