Servo - February 2009

■ FIGURE 4. Base of the system.

Table 1 - Parts List

Pendulum

Item Part# Qty Source Website

Optical Encoder M21AAS0DB03 1 DRC www.drc.com Encoder Base - 1 - - Hut HUB-02 1 Lynxmotion www.lynxmotion.com

Propeller 5x3 LP05030 2 APC Propellers www.apcprop.com

Motor 799C 2 Astroflight www.astroflight.com

Acrylic Rect. Bar 443501U.S.Pl astic www.usplastic.com

5. 5" x 1.4" x 0.225" Acrylic Rect. Bar 44350 1 U.S. Plastic www.usplastic.com

11" x 1.4" x 0.225" Alum. L Connector ASB-06 4 Lynxmotion www.lynxmotion.com Bracket 7/8-9 UNC Screw - 15 Any store -

1/2" Long 7/8-9 Nut - 15 Any store -

Electronics and Power Supply

DSP TMS320C6713 1 Texas www.ti.com

DSP Starter Kit Instruments E3631A 2 Agilent

commands. The motor and its velocity control can be purchased at

AstroFlight ( www.astroflight.com).

To measure the inclination of the pendulum, a two-channel quadrature encoder was attached to a Power Supply custom-made aluminum base as shown in Figure 4. The sensor we used was produced by DRC ( www.drc.com). Unfortunately, they do not produce it anymore. However, US Digital ( www.usdigital.com) produces similar and more modern products.

We powered the motors using a 12V, 1A power supply. The sensors and the motor controllers are connected to a TMS320C6713 DSP Starter Kit board. This project could be implemented using more simple and less expensive processing units, as long as they are capable of reading data from encoders and producing PWM signals.

www.home.agilent.com

integration, and derivation. Controllers using modern techniques make use of linear algebra and mathematical models of the robots being controlled. Analog computers are built with operational amplifier circuits, which makes it extremely hard to operate with vectors and matrices. In the following paragraphs, we describe how we implemented a classic digital PID controller on a DSP.

Controlling The Robot Using One Motor

Digital PID Controllers

A controller is run by an algorithm which makes decisions on how to manage the behavior of a plant or system, e.g., a robot. The control system theory is divided into classic controls and modern controls. Classic techniques may be implemented using analog or digital computers. However, modern techniques may only be implemented in digital computers. Classic techniques use basic mathematical operations like addition, subtraction, multiplication,

Motor1

As mentioned earlier, the objective of the system is to reach a given inclination set point angle. We want to achieve this as quickly as possible and with the least number of oscillations. To simplify and learn more about the behavior of the system, we first operated it with one motor. Figure 5 shows the control system's negative feedback loop we implemented. Later on, we'll describe how we added a second motor to try to increase the robot's performance.