Practical Approach to Rudder Control System for UAV using Low Cost MEMS Sensors

Transcription

1 Practical Approach to Rudder Control System for UAV using Low Cost MEMS Sensors Than Htike, Tin That Ngwe, and Yin Mon Myint Abstract Applying Micro Electromechanical Systems (MEMS) inertial sensors for the Guidance, Navigation and Control (GNC) of an autonomous Unmanned Aerial Vehicle (UAV) are an extremely challenging area. This paper presents a practical approach of applying a control system to control one of the control surfaces of UAV using MEMS inertial sensors, microcontroller and servo motor. In the paper, the control surface (rudder) is controlled in two modes; manual mode using a joystick and automotive mode using an accelerometer. The required yaw position of UAV is controlled manually and its stability is maintained by means of control system including an accelerometer. The whole control algorithm was implemented within a microcontroller. The future goals of this research are to incorporate more sensors to increase the level of autonomy for UAV operation. This paper addresses to design a very cheap and practical control system to approach modifying the UAV navigation system using low cost MEMS sensors. In this paper, we take advantage of commercial test equipment and have based our design on the basic ideas of sample platform. Keywords Microcontroller, Micro Electromechanical Systems, Servo Motor, Unmanned Aerial Vehicle. A I. INTRODUCTION N uninhabited air vehicle has found diverse applications for both civil and military missions. To achieve the stated mission, the vehicle needs to have a certain level of autonomy to maintain its stability following a desired path under embedded guidance, navigation and control algorithm. To meet the increasingly more stringent operation requirements, the UAVs rely less and less on the skill of the ground pilot and progressively more on the autonomous capabilities dictated by a reliable onboard computer system. Therefore, the objective of this control system is to deliver a control command to steer UAVs control surfaces and maintain them in stable condition. Unmanned air vehicles (UAVs) are attracting a large degree of interest in the navigation and control communities. Not only do UAVs provide excellent low-cost test beds for navigation system experiments, but their design and control facilitate the exploration of many exciting new research areas in control theory, ranging from low-level flight control algorithm design and mode switching experiments to high-level multiple aircraft coordinated mission planning. Autonomous navigation vehicles usually employ multiple sensors of various types. The technical challenges for small UAV systems are numerous since they will only be able to carry the smallest microprocessor systems and power supplies along with very lightweight and in expensive sensor systems. Authors are with Department of Electronic Engineering, Mandalay Technological University, Mandalay, Myanmar ( misterthanhtike@gmail.com, lwinlwinth@gmail.com). Fig. 1 Avionics system of UAV Normally, the architecture of UAV is as shown in Fig.1.The content of this paper is the former step of overall UAV project. In this system, firstly the orientation of UAV is placed to get the required yawing movement manually. The system takes the position of UAV along the Yaw axis until the auto mode of stability is switched. If it does, the control system senses the current coordination of UAV and adjusts the yawing movement to be stable in the required position. Although ADXL202 senses only rolling and pitching moment, we designed the model with yawing movement in the test bed. II. SYSTEM HARDWARE CONFIGURATION The overall system configuration is briefly represented in this section. Each component s characteristic is also explained. System architecture is as shown in Fig. 2. A. PIC Microcontroller The PIC 16F877 (Microchip Technology, Inc., microcontroller was chosen to obtain data from sensors, perform control calculations, and control the motors on the UAV. This microcontroller has a 25 MHz processor (the current compiler runs the processor at 20 MHz), 33 input/output (I/O) pins, (8K*14words) of Enhanced FLASH program memory,( 368*8bytes) of RAM, (256*8bytes) of data EEPROM. The PIC does not have an operating system and simply runs the program in its memory when it is turned on. 357

2 This PIC microcontroller has several hardware features that are very useful for use in a UAV and simplify the interfacing of sensors and motors with the microcontroller, such as an analog to digital converter (ADC), interrupts, timers, and capture/compare/pulse width modulation (CCP) channels. Fig. 4 Servos are controlled by 1-2 ms pulses Fig. 2 Block Diagram of System B. Servo Motor (F-S148) This kind of servo motor was chosen according to the following advantages. 1) High torque at all speeds 2) Capable of holding a static (no-motion)position 3) Able to reverse directions quickly 4) Able to accelerate and decelerate to reach a position or rate of speed quickly C. Servo Control The servo control board is a homemade prototype. The processor is PIC16F877 with 8 channel PWM signal out put. It can command 8 servos at the same time with RS-232 serial port. PWM signal is used extensively on DC servo control, such as the hobby model DC servo. A square wave is outputted 50 times per second. The width of the square wave decides the horn of the servo oscillating angle, and the wave width is described according to the continuous time. When the width of the square wave equals 1.5 millisecond, the horn of the servo keeps on neutral position, 45degree angle. The width of square will change from 1 to 2 millisecond, and the horn of servo will rotate amount 0~90 degree angle as shown in Fig. 4. D. MEMS Sensor (Accelerometer) A dual-axis accelerometer from Analog Devices, Inc. (ADXL202) was interfaced with the PIC microcontroller. The accelerometer is capable of sensing accelerations of+/- 2 g s in two perpendicular axes. The sensor chip produces a digital pulse proportional to it s along two axes by using a duty cycle modulator that is included as part of the sensor chip. In this paper, only the signal sensed in one channel (x-axis) is used to represent the state of yaw angle although both two channels are in use in calibration. E. Homemade Joystick The design of homemade joystick is made to be able to control 4 servo motors simultaneously. In this paper, joystick is to manage a rudder of small UAV model to get required yawing movement. Physical model is designed to be able to adjust the range of angle of rudder with response to the range of angle of joystick. F. Small Model UAV To test the control system, a small model UAV was designed as a test bed. The range of angle of rudder was designed to have a range of 30 degree. III. DECODING ALGORITHM The signal from both axes on the accelerometer consists of a square wave, where the pulse length is proportional to the acceleration sensed on that axis. If the base period of the signal is T2, an acceleration of 0g means that the length of the pulse T1 is T2=2. A maximum positive acceleration means that T1 = T2, and a maximum negative acceleration is represented by T1 = 0. The ADXL202 can measure accelerations in the range +/- 2g. The duty cycle change per g (where 1g = 9:81m=s2) is 12:5%. A straight-forward way of measuring the acceleration sensed by the two axes x and y is outlined below: 1) Start the timer at the rising edge of the first channel, x. 2) Stop it at the falling edge. This gives T1x. 3) Start the timer at the rising edge of the second channel, y. 4) Stop it at the falling edge; and one gets T1y. 5) Compare these values to T2. Repeat. This is straight-forward, but it has the disadvantage of being slow. One can only get a sample of the acceleration of both channels once every three cycles (3*T2). One wastes cycles waiting for the next channel's rising edge, as shown in Fig. 5. There is fortunately a solution to this problem. The timing of the pulse output is handled by a triangle wave in the ADXL202. This means that the midpoint of T1x and T1y coincide in time. This means that the time between the midpoints of T1x and T1y is equal to T2. Since we know the value of T2, we can use this fact to decode the acceleration in a smaller timeframe than before. In short, we use the following algorithm: Fig. 3 Servo motor 358

3 practice the zero g output and the sensitivity of the ADXL202 vary somewhat from device to device (see the data sheet for details). So this formula can only be used for low accuracy measurements. For higher accuracy measurements we have to substitute the actual offset and scale values. Fig. 5 Basic Decode Technique for the ADXL202 Fig. 6 High speed Decoding Technique for the ADXL202 1) Start the timer at the rising edge, Ta of the X channel. 2) Stop the timer at the falling edge, Tb. By dentition, T1x is now equal to Tb-Ta. 3) Repeat for the Y channel, and get T1y = Td - Tc. 4) As T2x = T2y, get Te- Ta = Tg- Tf, and after substitution T2 = Td (1) The advantages of this method are twofold: we acquire one sample of acceleration from both axes every two T2 cycles (compared to three cycles in the previous algorithm); and T2 is calculated only once for both axes. IV. CALCULATION OF ACCELERATION OUTPUT Acceleration experienced by the ADXL202 may be calculated by the following formula: Acceleration (in g) = As outlined in the data sheet, the nominal duty cycle output of the ADXL202 is 50% at zero g and 12.5% duty cycle change per g. Therefore to calculate acceleration from the duty cycle: (2) Acceleration (in g) = (3) When the zero g duty cycle output of the ADXL202 is other than 50%, and/or the duty cycle changes more or less than 12.5% per g, the acceleration calculation will be inaccurate. In V. CALCULATION OF ADXL202 When in calibrate mode the accelerometer must be level, with the X and Y axis horizontal to the earth so both axis experience 0 g. When the microcontroller is told to calibrate, the microcontroller reads the duty cycle output (T1) and period (T2) of the accelerometer from each axis. Several readings of T1 and T2 may be averaged to improve accuracy. These values are stored as calibration constants, and retained for use in calculating the acceleration after calibration. A scale factor (K) used to scale the acceleration output into an n-bit word should also be calculated. The calibration constants for each axis consist of the following: 1) T2cal = the value of T2 during calibration. T2cal must be stored as the value of T2 does drift over temperature and has jitter. 2) Zcal = the value of T1 during calibration. 3) Bit scale factor. You choose the bit scale factor to determine the resolution (in bits) of the acceleration calculation. 4) K = [4 * (T2cal * bit scale factor) / T2cal] Note that K need only be calculated once for the two axes. VI. CALCULATION OF ACCELERATION FROM DUTY CYCLE Once the calibration constants are known, only two formulas are required for the calculation of acceleration. They are: Zactual = (4) Where T2actual is the current measurement of T2. This formula corrects the zero g value for changes in T2 due to drift or jitter Acceleration = (5) VII. CALCULATION OF ANGLE When the angle is varied along the sensitive X-axis and Y- axis, the acceleration vector changes and the ADXL202 responds by changing the duty cycle outputs. The angle is defined by the following equation: θ = arcsin[ (V(out)-V(zero g)) / (1g x Scale factor(v/g)) ] (6) VIII. SYSTEM SOFTWARE DESIGN A. System Integration Fig. 7 is a flow chart of the program of the system. Firstly, all of the required initialization was set up. Especially, the program declares the setting of PWM mode(ccp),timer and A/D converter. 359

4 1) PWM mode in CCP Channel Since servo motor is controlled by means of managing PWM, the features of CCP (Capture/Compare/Pulse Width Modulation) play in important role. The period of the PWM is set up using the following formula: PWM period = <(PR2)+1>*4*Tosc*(TMR2 Prescale value) (7) The PWM duty cycle is specified by writing to the CCPR1L register and to the CCP1CON<5:4>.The overall meaning of the control system is to calculate the required duty cycle to control the servo motor. 2) Timer and Counter Tmer2 is configured to produce PWM with the required frequency. According to the nature of servo motor, the frequency is set up to be 50Hz.To record the output of ADXL202, count configuration is set up. 3) A/D converter The 10 bit analog to digital converter on the PIC microcontroller is used to convert the signal representing the joystick movement to an integer that could be used by the microcontroller. The resolution is significant enough to do the main idea. B. Control Program Fig. 7 Flowchart of the program As shown in Fig. 7, when the program is started, the control surface waits to ensure that it is manually controlled initializing configures. The control algorithm reads the joystick movement and computes the values that are loaded to control the servo motor to get required yawing movement. When the controller senses the stability request, it reads the current position of UAV and then it computes the real time tilt angle of UAV as the error of position. Finally, using some control algorithms it recovers the state of yaw position in stable. Fig.8 shows the designed control system in test bed. C. Display Unit We used Hyperlink software as the monitor. The unit including RS-232 processes this nature. The program displays the state of yaw angles controlled by joystick, sensed by ADXL202 and recovered by the system. IX. SUMMARY In this paper, our work addresses practical approach to control yawing movement in two modes. The control system has been designed with an emphasis on using inexpensive MEMS sensor and simple algorithm. Although the stability augmentation has been complemented in test-bed, more complete system will be conducted in the future work. ACKNOWLEDGMENT The kindness and help of Dr. Zaw Min Naing and Dr. Yin Mon Myint, Department of Electronic Engineering, are gratefully acknowledged. I also wish to thank my members of UAV project for their work and warmly friendship. Especially, I would like to express my special thank to my parents for their noble support and encouragement. REFERENCES [1] Jung Soon Jang and Darren Liccardo, Automation of small UAVs using a low cost MEMS Sensor and Embedded computing Platform, Crassbow Technology, Inc., San Jose, California. [2] Hoffmann, Gl, Rajnarayan, D.G., Waslander, S.L.,Dostal, D., Jang, J. S., and Tomlin, C. J., The Stanford Testbed of Autonomous Rotorcraft for Multi-Agent Control(STARMAC), Salt Lake City, UT, 23rdDigital Avionics System Conference,November,2004. [3] Ian Schworer, Navigation and Control of an Autonomous Vehicle, Blacksburg, Virginia, [4] Ari Yosef Benbasat, A Inertial Measurement Unit for User Interfaces, Massachusetts Institute of Technology, [5] Reed Diefert Christiansen, Design of an Autopilot for Small Unmanned Aerial Vehicles, Brigham Young University, [6] Rebecca Jane Dailey, An Automation Approach to Guiding an Air Vehicle Through an Obstacle Field, Holt, Massachusetts Institute of Technology, [7] Carlo Canetta,Jonathan Chin,Sevan Mehrabian, Ludguier Montejo, Hendrik Thompson, Quad-rotor Unmanned Aerial Vehicle, Final Report,May 2, [8] Rodger Richey, Measure Tilt Using PIC 16F84A & ADXL202, Microchip Technology Inc,1999. [9] Paul Y. Montgomery, Carrier Differential GPS as a Sensor for Automatic Control, June [10] Kenin J. Walchko, Low Cost Inertial Navigation: Learning to Integrate Noise and Find Your Way, University of Florida, [11] Alexandros Skafidas, Microcontroller Systems for a UAV, 4th December [12] Dr K.C. Wong, UAV Design Activities in a University Environment, University of Sydney,

5 Fig. 8 Control System in test bed 361