Not : Daha önce yazdığım proje yazısının birebir olmasa da çevirisi sayılabilecek bir yazı hazırladım. This is a short summary of our (me and my friend Umut Yuksekbas) graduation project. Sorry for the grammatical mistakes.

Computer & Microcontroller Based Closed Loop Robot Position Control System

Abstract : Robot systems scope out the problems fast and sensitive, don’t affected the factors which destructive of health for humans. For this reason they are needed at industrial areas. Robots are used for precise applications in a wide area such as automotive industry, medicine, nuclear power stations and biomedical. They can be used in critical applications like sensitive surgical operations. That’s commonly utilized as a mechanical arm for handicapped ones. Our project’s subjects are automatic control systems and position control of a robotic arm with the coordinates that send from computer. We intended to design a closed loop control system of a mechatronic system. Manipulator system is a multi-input multi-output (MIMO) system and each joint is fed by different input signals for position check. Dynamic model of the manipulator system is not linear and contains interactive terms between joints. We have designed a 6-degree-of-freedom robot arm and a control system. DC motors on the joints provide the movement of the manipulator. We used L298 IC which contains H-Bridges for driving the DC motors and PIC 16F877 as microcontroller. Max232 serial interface sets up the communication between computer and microcontroller. We developed the position control algorithm in MPLAB IDE by using the Hitech PICC Compiler. Microcontroller takes the analog feedback signals from the potentiometers placed on joints then convert them to digital signals by built-in ADC modules. The PI control system uses the analog signals in order to calculate the optimum PI control parameters then provides the target movement with minimum error.

SYSTEM COMPONENTS

Hardware Component

PIC 16F877-20MHz contains 8 ADC channel and 2 PWM module and L298 contains two H-Bridge in it. This driver is very robust and is able to work with 46 V source voltage and 4 A source current. In this project  15 V source voltage and 0.5 A max current are utilized.

Block Diagram

Position data which transmitted from computer come to the PIC’s serial port over the Max232 serial interface. After that, the ADC module reads and converts the analog values and transmits signals to the motor drivers for holding the arm in target position. The algorithm determines the transaction direction for decreasing the error and sends the signals. It breaks the motors when the arm is on the target position. Microcontroller checks all of the joint positions repeatedly and if it senses a deviation on the joint, brings the joint back to the target position. The algorithm changes the motor voltages directly proportional to the error

Schematics

PCB Design

A 2-layer circuit board was designed and assembled. L298 ICs can be seen on the right side of the first image. The sockets from top to bottom;

• Power
• Analog Feedbacks (From joint pots)
• Motors

PCB 1

PCB 2

Software Component

We developed the microcontroller code by using HITECH PICC in MPLAB IDE. The P.I. Control System on the microcontroller side, calculates the error by differentiating the feedback values which read from ADC module and the expected value, continuously for each joint. This error value multiplied by Kp constant. All these error values summed and multiplied by Ki constant. The control signal which derived by summing the proportional and integral control signals, limited in a desirable range.

Control Signal

This calculation repeats for each joint over and over again for providing the position control. Ziegler-Nichols method were used for designing the control system, because it was not possible to get the technical data about mechanical system for deriving the mathematical model.

Controller Design

D and I constants of PID controller let by zero and, the P constant were increased until the system oscillates. The P constant on that oscillation time called Kc and the oscillation frequency called Tc.

Oscillation

The optimum P.I. Control parameters were derived by using the Ziegler-Nichols method. The system response is below mentioned.

System Response

Graphical User Interface Design

Graphical User Interface (GUI) was developed by using QT Designer and Kdevelop Integrated Development Environment in C++ programming language. The GUI design can be seen below. QT libraries are suitable for creating platform independent code. By this way, the GUI can be compiled and run under various platforms, such as Unix, Linux, FreeBSD, MacOS and Windows.

GUI

The coordinate data transmitted over the USB bus by using a USB/Serial converter with a 57600 baud rate. Control System variables can be changed or modified with the GUI in realtime. Serial port device, baud rate and control system parameters can be selected and changed by using the GUI.

GUI Ayarlari

Conclusion

The system that we developed can do a position control in Cartesian Coordinate System. It keeps its position unless the operator will give a new coordinate, and can perform a coordinate change in a desirable time. I believe that this project is providing a useful design for industrial areas such as production and test, biomedical, security, situations that harmful for human, and academic study. The future plan for my study is making two robotic arms work together and in cooperation.

Bugün süper bi özelliğimi keşfettim. 2 haftada herkesi sular seller gibi Japonca bildiğime inandırmışım, nasıl becerdiysem. Bu inandırıcılığımı başka işlerde de kullanırsam köşeyi dönebilirim. Nasıl oluyor da oluyor anlamıyorum. Bugün Tokyo ve Kyushu Üniversitelerinden seminer için gelen 2 profesörün sunumlarını dinlemek mecburiyetinde bırakıldım. Yaklaşık üç buçuk saat sürdü. Sıkıntı ve uyku birbirine karıştı. Arada yakalayabildiğim – önceden duymuşluğum olan – birkaç kelimenin anlamı neydi diye düşünürken o kısım bitip diğerine çoktan geçiyordu zaten. Okuma yazma bilmeyen çocuğun kitabı alıp resimlerine bakması gibi slaytlardaki resimleri inceledim saatlerce. Pek de özenerek hazırlanmıs slaytlar değildi. Hal bu ki ne de güzel çalışmalara benziyordu anlattıkları. İnandırıcılığım tavan yapmış olacak ki, çıkışta da seminerde öğrendiklerim hakkında rapor yazılması gerektiğini – tabi yine japonca olarak – söylediler. Bilenlere sordum da ordan öğrendim öyle söylediklerini. Yoksa bildiğimden değil. Şimdi rapora dizeceğim incileri düşünüyorum. Hangi resimden başlasam anlatmaya acaba? Bir tanesini net hatırlıyorum. Bir yerli (Japon) ve bir yabancının (Ben) kanji okuduğu sıradaki beyin fonksiyonlarını karşılaştırmalı gösteren hareketli bir grafikti. Epey de hatılıyorum bakın, o kadar da boş boş bakmamışım yani. Bu şekilde duyarak öğrenmeye devam edersem önümüzdeki yıl bu vakitlerde おかあさん(anne), おとうさん (baba) diyor olurum herhalde. Bi yandan Japonca dersi de alıyorum da kısa zamanda yırtarım inşallah. Herşey kuralına uygun olsun diye önümüzdeki hafta Salı günü anne-baba deyip bir üst kademeye (mama) geçmeyi planlıyorum. Bu arada araştırma konumla ilgili detaylar da hafiften gün yüzüne çıkmaya başlıyor. Henüz kendisini göremediğim bir robor kol ile haşır neşir olacakmışım. En güzel yanı da embedded Real-Time Linux kullanıyor olması. Kesin olmamakla beraber robotun acı çekmesini sağlayacak bir proje olması muhtemel. Detaylar önümüzdeki zamanlarda ortaya çıkacaktır.