This is to certify that I am responsible for the work submitted in this project, that the original work is my own, except as indicated in the references and acknowledgments, and that the original work contained herein has not been performed or performed by unspecified sources or persons. The robot is programmed with Linux as OS and C language to execute the algorithm. Experiments are conducted to become familiar with the tools used in this project.
Meanwhile, for CMUcam2, it is interfaced with the PC and basic calibration is done before the picture is taken. I would like to express my highest gratitude to the people who helped me in completing this report. He also shares his experiences which were really helpful in doing this final year project.
He always has a solution for my problem and also motivates me to complete this project. Yasin in building the robot model. Not to forget my fellow colleagues, Khairul Ariffin, Alfred, Rasid, and Zulhilmi for their motivation and encouragement they gave me to go further in my life.
- Background study
- Problem Statement
- Objective
- Scope of Study
Raspberry Pi is 3.3V, therefore there will be some problems when controlling motors such as DC motors or servo motors. The first problem is to get the robot to work as close to real time as possible. Because image processing is involved, there may be some delay in the robot's decision-making.
A proper method should be implemented to ensure that the image processing does not take much time and can work in RTOS. This is because the Raspberry Pi is still new and not many people have tried to connect it to other devices, in this case the Raspberry Pi with CMUcam. Theoretically, the interface is performed via serial communication and the CMUcam should respond to the command sent by Raspberry Pi.
The most important thing is that the price of this robot should be low compared to the available robot in the market that offers the same functionality. The goal of this project is to build a cheap small mobile robot based on Raspberry Pi and CMUcam2.
Image processing on Robot
Real-Time Operating System (RTOS)
Some of the other RTOSs available are µITRON, EmbOS, Each of these RTOS has different features in terms of system management, interrupt management, memory management and many more [10]. In ref [10] they compared all available RTOSs and found that each of the RTOS has its own advantages and disadvantages.
The advantage of Linux is that it is open source, has many software resources and supports multithread, user and process. The reason to choose Linux OS is because it can manage processing and work in real time.
Decision Making Techniques
- Research Methodology and Project Activities
- Experimental Procedures/Approach
- Programming Raspberry Pi
- CMUcam2
- Mobil robot movement
- Interfacing
- Key Milestones
- Gantt Chart
The method used to track the object is first by capturing the image and transferring it to the Raspberry Pi (Microprocessor). Once the image is obtained, it is processed and analyzed to see if the color of the object matches the color of what had been programmed. Basically, the area that the object could be is selected first and the color of that zone is obtained when the object is found within the zone.
Then the Raspberry Pi sends signals to the servo motor to move towards the object. Once the data is acquired, it is sent to Raspberry Pi to process the data. The Raspberry then sends the command to the motor to make one of six movements.
Firstly, Raspberry Pi needs to be programmed to communicate with CMUcam or other devices. Sufficient time should be spent testing the device and performing simple tasks to become familiar with the built-in operating system programming. Furthermore, this device is considered still new on the market and will therefore require extensive testing on the board to ensure that it is really suitable to be used in this project.
Therefore, acquiring the image data under different lighting conditions will be helpful when it comes to the algorithm to detect the color of the object. The robot's movement is about how the robot will respond to the image and obstacle that the robot detects. The robot will have to decide whether to go forward, backward, left or right depending on the object it is tracking.
An algorithm must be created to ensure that the mobile robot is able to move according to the object movement. Once the system is in place, it is tested to get the accuracy of tracking and movement of robot in different environmental conditions. Test the operation of the system to get the accuracy of detection and the movement of robot in different environmental conditions.
- Setting up The Raspberry PI
- Communicating the raspberry pi with the outside world
- Simple Led Blinking
- Serial Communication
- CMUcam
- Robot Platform
- Motor Driver
This is the window in which it will allow the user to do programming and also allow the user to make the raspberry pi to communicate with the outside world. There are many methods that can be used to communicate the Raspberry Pi with the outside world. Pi Face is a device that can be connected directly to the Raspberry Pi before it is connected to the outside world.
In the future, a different approach will be used to connect the Raspberry Pi to the outside world. Where the user has access to the GPIO pins on the 26 pin of the Raspberry Figure 12 Interfacing with Raspberry Pi and Pi. The serial communication in Raspberry Pi is done by using the UART pin available on the GPIO Pi on pin 14 and on pin 15 as shown in the image below.
To access the raspberry pi's UART, it must be configured first. To run minicom, it must first be installed on the Raspberry Pi by typing sudo apt-get install minicom. Request the use of the serial port by typing /ttyAMA0/ and comment it out by placing it at the beginning of the text and reboot.
This is because the Raspberry Pi runs at 3.3v while the RS-232 connection can have up to a potential of 15V running through it. As mentioned above, the Raspberry Pi runs at 3.3V, so the logic switch will be useful to drive a higher voltage circuit. The image below shows the communication image and also the serial communication circuit.
In this project, the center of the image is obtained and refreshed every few seconds. To test the function of the camera, the CMUcam2 GUI can be used to test the camera and take sample photos. For the test purpose, the CMUcam is attached to the servo motor and controlled through the serial port.
Even if the movement of the camera to the image is not so smooth, but the camera is still able to detect the color of the object. The space at the top of the robot will be used to place the Raspberry Pi and CMUcam.
Xinqiang, "The design of intelligent robot based on embedded system," in Advanced Mechatronic Systems (ICAMechS), 2011 International Conference on, 2011, pp. Houshangi, "A vision-based autonomous lane following system for a mobile robot," in Systems , Man and Cybernetics, 2009. 34;Mobile Robot Vision Tracking System Using Unscented Kalman Filter,” in System Integration (SII), 2011 IEEE/SICE International Symposium on, 2011, pp.
Nourbakhsh, "Second Generation Low-Cost Embedded Color Vision System," in Computer Vision and Pattern Recognition - Workshop, 2005. Junzhi, "Embedded Visual System for Robotic Fish Navigation," in Computational Applications and System Modeling (ICCASM), 2010 International Conference on, 2010, p. Madhavi, "Wireless Vision Based Moving Object Tracking Robot Through Perceptible Color Space," in Emerging Trends in Robotics and Communication Technologies (INTERACT), 2010 International Conference on, 2010, p.
Rover, "Enhancing students learning with hands-on RTOS development in real-time systems course," i Frontiers in Education Conference, 2008. Tianmiao, "A novel multi-OS architecture for robot application," in Robotics and Biomimetics (ROBIO), 2011 IEEE International Conference on, 2011, s. Ren-fang, "Research of key technologys for embedded Linux based on ARM," i Computer Application and System Modeling (ICCASM), 2010 International Conference on, 2010, pp.
O_NOCTTY - When set and path identify a terminal device, open() should // not make the terminal device the controlling terminal //for the process.
