SMART BIRD TEAM UAS JOURNAL PAPER

Transcription

1 SMART BIRD TEAM UAS JOURNAL PAPER 2010 AUVSI STUDENT COMPETITION MARYLAND ECOLE POLYTECHNIQUE DE MONTREAL

2 Summary 1 Introduction Requirements of the competition System Design Design Approach System Overview Auto-pilot development Accelerometer and gyro meter Control Section Navigation Navigation Simulation Ground Station Primary flight display Map Data Video Mechanical General Considerations Radio control system Recovery system Design of the recovery system Required parachute area Inventory of required equipment Power and control system Conclusion Acknowledgments References Appendices... 16

3 ABSTRACT This paper discusses how undergraduate students have put together all the information and knowledge they have, to design and to implement an Unmanned Aircraft System (UAS) for AUVSI The aircraft may be capable of performing a predefined path, following different waypoints. On other side, that aircraft may also be capable to detect and identify precisely some alphanumeric targets and pop-ups while in flight. It is our first participation to this competition and we are going with an Avistar powered by an electrical motor Axi 2826/12 (Gold Line) and equipped with a KX191 color and night mode CCD camera, an AC M-01 modem (designed by Laird Technologies) and our PCB (printed Board Circuit). Our scope is to design an autonomous aircraft consisting of a PCB, designed by the team, capable of controlling and navigating. It is also to design a performing ground station which helps us following the aircraft movement in real time, the video sent by its on-board camera and seeing targets form and color.

4 1 Introduction UAS take, nowadays, a great importance among civil world and its applications while it was mainly considered as nation s armies affair. It can be used to detect people under a block when there is a seism, or people blocked in a building in fire. It can also be used for a specific action where human life can be in danger like sending a packet form a point of the globe to another. Our team goal is to develop its own PCB and connect to it many sensors like GPS, gyro meter XY, gyro meter Z or even a speed sensor, an SD Card and so on depending on the application we want. In the case of this competition we chose to use only four sensors that will be described further. Comprised of 20 undergraduate members, the PCB developed by the team (6 members) is able to take 6 sensors, to autopilot the aircraft and to do the navigation by waypoints. On the other side, ground station team members developed their own application made of 4 windows (PFD, flight Data, Video and Map) allowing them to have a full control of the aircraft, its position, its movements and to communicate with it. 2 Requirements of the competition Key Performance Parameters Threshold Objective Autonomy Imagery Target Location Mission time(1) In-flight re-tasking During way point navigation and area search Identify any two target characteristics (shape, background color, orientation, alphanumeric, and alphanumeric color) Determine target location ddd.mm.ssss within 250 ft Less than 40 minutes total Imagery/location/identification provided at mission conclusion Add a fly to way point All phases of flight, including takeoff and landing Identify all five target characteristics Determine target location within 50 ft 20 minutes Imagery/location/identification provided in real time Adjust search area Table 1 : Competition requirements

5 3 System Design 3.1 Design Approach In order to successfully complete the project, the team was divided into three subteams. They worked under the coordination of a team lead and their goal is to harmonize elements and integrate them together. The three sub-groups are divided according to disciplines that affect the project and taking also into account student s specialisation. The three teams are: Mechanical, Control and Ground Station. Mechanical team is responsible for the mechanical assembly of the aircraft, designer of the system security and backup, and fly the aircraft. Control team, the most important, is itself divided into three parts that are the control, navigation and aircraft embedded system called PCB. It deals with control and stabilizes the aircraft by designing control loops and the controller compatible with the UAV. It is also responsible for programming control loops on the system board after designing the embedded and the electrical systems. Navigation sub-team deals will build of the navigation algorithms using the waypoints (GPS coordinate). Last team, Ground Station, scope was to the development of a GUI based on the algorithms of detection of colors and shapes. 5

6 Figure 1 : Task Division 3.2 System Overview The PCB put inside the aircraft is capable of controlling the aircraft automatically. The algorithms on the PCB implement the control of two micro controllers. The map of the PCB consists of two microcontrollers who communicate by UART. On one of the microcontrollers, we have acquisition of different sensors such as the accelerometer, Gyro XY, barometer, compass and the modem that communicates with 6

7 the ground station by Wi-Fi (RS285). The modem is capable of communicating with a frequency in MHz and MHZ. On the second microcontroller, we find that the sensors are gyros Z, the GPS, the outputs of four servo motors. Three of those servo motors serve to control the flight surfaces and the last one is used to control the rotation of the camera. The transmission of the video is on a separate frame, as well as the data sent to the ground station. It uses 900 MHz FM Wireless 500mw Video Transmitter to transmit video from the camera to the ground base. The PCB consists of two micro-controllers because the sensors used have different voltages wide range of public. The sensors connected to the first microcontroller are on voltage 3.3 V while those connected on the second microcontroller operate at 5 V. We use the Max3002 to convert data between micro-controllers since they communicate through UART. For competition, we use only 4 sensors: the accelerometer, the XY gyro, the Z Gyro and the GPS. We use also a modem allowing us to communicate with the Ground Station. 4 Auto-pilot development 4.1 Accelerometer and gyro meter Frequent Flight parameters used to realize the navigation and control of an aircraft are: angles of roll and pitch, and altitude. Those parameters are determined using a GPS and various sensors such as the 4 listed above. The type of the accelerometer used is IMU and it provides acceleration values of the aircraft in three directions (X, Y and Z). The accelerometer and gyroscope are used to evaluate all the angles of roll and pitch of the aircraft. For this we use an additional filter that takes into account the accelerometer data to calculate the angles at low frequency and the gyro data at higher frequency. 4.2 Control We use a proportional integral controller for achieving the subjugation of the UAV. By doing that, we faced a major problem: we do not have a very viable model simulation of our device because we run out of time to do so. We then made open-loop tests using the remote to rise and angle values and outline transfer functions, using PWM signals generated, in order to achieve a safe control of the aircraft. The servo parameters are the angles of roll and pitch, rudder and throttle. The control is done using the dspic30f microcontroller, which generates PWM signals suitable for 4 servos. These actuators are then connected to the flight surfaces to allow the stabilization of the device. 7

8 4.3 Section Navigation Navigation The task of the navigation team is to program the navigation of the drone so that it can follows a specify path or change path when in flight. For this purpose, we decide to use the Open Source Project Ardu pilot. Our goal was to analyse their code and retrieve functions we needed to implement the navigation. The first thing we did was to compile their project on Visual C++ and to try to execute it. We got some problems with compilation because we didn t have the same IDE (Interface Development Environment) as people who developed the project. So we decide to modify the project and we put functions that have the same goal in the same file to avoid compiling issues. Now, the functions are separated in two files, one for the navigation and another for the control. It is easier for us to work like that since there were two sub-teams (control and navigation) Simulation To be able to test the code we developed, we decided to use Virtual Reality Toolbox in MATLAB / SIMULINK (Thanks to AUVSI and his partner MATLAB who provided us with that). We developed our code in C; in Simulink there is a tool named Legacy Code Tool that allow to integrate functions write in C into MATLAB / SIMULINK model. Using an airplane model of our drone in Virtual Reality Toolbox, we could test the code we implemented. So far we have found a model for the F-14 flight combat and another model where you can simulate by introducing position and speed versus time. These models are not suitable for our task and they were used just for some rapid tests. Finally we worked with a model, near the one developed by a part of our team. 5 Ground Station The ground station application has been developed in C#. We are using Microsoft.Net Framework. We also use Direct3D for graphic render and AForge.Net library for image acquisition and processing. For the communication with the UAV, we are using a wireless serial (RS232) modem from the company Laird. Our application has been separated in fours panels. The following image is a screenshot of our application: 8

9 Figure 2: Ground Station 5.1 Primary flight display In the first panel, we designed a primary flight Display (PFD) with Direct3D graphic library. On the center of the PFD, there is an altitude indicator which gives the information about the UAV s pitch and roll. This instrument also gives the information about the orientation of the UAV with the horizon. To the left of the altitude indicator, there is an airspeed indicator that displays the speed of the UAV. To the right, there is an altitude indicator that displays the altitude of the UAV. To the right of the altitude indicator, there is a vertical speed indicator that indicates if the UAV is ascending or descending. The heading indicator is placed at the bottom of the altitude indicator and shows the magnetic heading of the UAV. 5.2 Map The second panel is used to draw a map. A small blue square indicates the starting position of the UAV. Subsequently a blue line links the previous GPS coordinates of the UAV to the present ones. This gives the path taken by the UAV from its point of departure. Note that the Map works without internet connection. An image of the area to be over flown by the UAV is preloaded with the application. 5.3 Data The third panel is used to write all the information of the UAV like its position, its speed, its altitude This panel indicates the current mode of the application (acquisition or simulation). In the acquisition mode, the UAV sends data to the application and in the simulation mode, the application works with the keyboard. 9

10 5.4 Video The last panel is used to display the video on the screen. The play button launches the video and a context menu is available to stop/restart the video. We used the KX191 camera. Our choice was mainly governed by its light weight and high resolution feature. The machine vision system entirely works with Aforge.NET framework 1. Though we tried to work with other open source ones, Aforge.NET presented the best advantages including a set of libraries for video and image processing as well as some computation algorithms. Therefore, we used the Graham algorithm 2 and the neural network solution to find the targets. The detected target s characteristics such as shape, location, shape s color, alpha numeric, alpha s color and orientation are then written in an Excel board to be saved. Following screenshot shows some detection testing. Figure 3: Target detection test Graham s algorithm is a method of the convex hull of a finite set of points. 10

11 6 Mechanical 6.1 General Considerations First of all, we start to mount the plane and specially the part that supports the engine. The goal was to ensure the stability of this part of the plane to avoid the impact of the vibrations of the engine. After that we decide to do a flight test so that to check the power of the new electrical engine, and also to check the general ability of the plane to fly. But before the flight test, we ensured that the plane was balanced and its structural integrity was not engaged. We calibrated the radio controller and checked that all the flight controls are working well and in good conditions. The test has been done and we ve noticed that everything works fine. But we had some parts of the plane which needed to be repaired due to a hard landing. These parts were essentially the plastic support of the nose landing gear and the support plate of the engine. 6.2 Radio control system We needed a new radio since our previous one Futaba 4ch was missing 2 channel for our requirements (Parachute, Computer / Pilot control) and was not a programming Transmitter. We chose the Futaba 6EX 2.4GHz transmitter with the included receiver and equipments. Here are some reasons why we stop our choice on this radio: Futaba is top brand in market, Futaba specifications and documentations can be found online, Smart Bird team used Futaba before, Previous equipment is compatible with Futaba 6 channels Transmitter. This radio control is used to control: ailerons, rudder, elevator, throttle, On/Off parachute, Pilot Control/Computer Control. 6.3 Recovery system In case of signal loss or a major problem during the flight (Eg: fire), a recovery of the UAV must be provided to ensure its safe landing. With this in mind, a parachute system must be designed. During the previous term, the following steps of the design were made: 11

12 6.3.1 Design of the recovery system As it can be seen in the figures above, the recovery system design is fairly simple. The device is actuated by a servomotor, which cause the opening of the bottle of helium, through a striker. The balloon attached to the bottle then fills gas and propel the parachute. At the end of propulsion, the balloon bursts. The design is now completed, but it remains to determine the position of the device relative to the UAV, and the size of all its components Required parachute area To calculate the parachute area needed, a formula taken from the book «l exploration spatiale et ses techniques» of Bertrand Manuali was used. This formula uses the speed of descent of the UAV, and its total weight. S = m.g k.v2 rms Where: m: approximated total mass of the drone in kg g: acceleration gravity m/s² V: speed of descent of the UAV (m/s) K: aerodynamic coefficient S: surface opposing air m2 Numerical application: m = 2.8kg; g = 9.81m/s²; k = 1.2 and V = 2.5m / s, S = 3.67m². 12

13 6.3.3 Inventory of required equipment The tooling required to manufacture the system is summarized in the figures presented above. These include: A bottle of helium, A balloon up to a volume of at least 1 m³, A parachute fabric of an area of 3.7 m 2, A servo-motor, An extension to link the servomotor to the striker, Aluminum foils to hold all the components of the device. With a total approximate weight of 7lb we decide to go with an 8000 map battery, which will allow us to fly up to 30 min. We haven t done the test of the parachute yet but we re expecting another solution if the first doesn t work. The second solution consists in flying with the parachute during the whole flight. But this will increase the drag and reduce the autonomy. We have to make a trade off. 7 Power and control system Since the powering of our plane is electrical, power supply was a critical task to achieve. Main driving force of the UAS is a brushless DC (BLDC) motor. It is driven by a speed controller which control signal comes either from a microcontroller, or from the RC receiver. A PCB has been designed that supports two (2) microcontrollers, a 3.3V-powered one (core 1) and another 5V powered (core 2). Both are fed by the main battery using voltage regulators, MIC29310 for 3.3V and LD1085xx for 5V). Core 1 is dedicated to sensor acquisition and communication with the ground station. Core 2 realises mechanical control: switch between manual and auto pilot, sending PWM signals to servos and control signal to the speed regulator when in autopilot. Also, since GPS and Gyro-Z are 5V-powered, they are connected to core 2, which sends their signals to the other microcontroller. Communication between two cores is achieved using a MAX3002 transducer. Following table summarizes connection of sensors and servos to the cores and Figure 2 presents an overview of our systems and their connections. 13

14 Core 1 Core 2 Voltage Level 3.3V 5V Input signals Accelerometer GPS Barometer (through MAX232) Compass Gyro Z Gyro XY SD card Output signals N/A MUX selection signal 3 Servos and speed controller (through MUX) Bidirectional signals Core 2 (through MAX3002) Core 1 (through MAX3002) Modem Table 2: Summary of components' connection to the microcontrollers Figure 4 : UAS System Achitecture Components used for power and control Accelerometer: LIS3LV02DQ Barometer: SCP1000-D11 Battery Compass: PNI V2Xe 14

15 Gyro XY: IDG300 Gyro Z: ADXRS301 Microcontrollers: DSPIC30F4013 Modem: AC4490 Motor : AXI2826/12 Multiplexer : 74HC257D 3.3V regulator : MIC V regulator : LD1085xx Speed controller: JETI advance 40 opto plus Traducers: MAX 232 and MAX Conclusion For participating to this competition, we got so many issues that we are very happy to be at this stage of the competition. We tried and developed our own PCB, equipped with different sensors. We took a lot of time to program each function of the microcontroller. It was very challenging to do all this design by ourselves and to create our own ground station. It took a lot of time of work but at the end we are very happy of the result we got. We even run out of time at the end because we wanted to explore some innovative things to resolve our issues. Finally we took the simple way and assure a safe solution. It sure that, returning back from this competition, we will do a debriefing and think about the way to improve our systems. For example, a challenge for the next year will be to power some of our systems like the PCB with solar energy. It is on standby for now. 9 Acknowledgments We sincerely acknowledge teachers, professors for all their technical support on control and navigation. We also want to acknowledge the electrical department of Polytechnic College of Montreal which allows us to develop all that system by giving us all the administrative support: Richard Hurteau, Electrical Department. Lahcen Saydy, Electrical Department. David Saucié, Electrical Department. François Morin, Technical societies head at Polytechnic of Montreal Nanorobotic Department for their support for the PCB. 15

16 10 References 1) Dspic30f4013 reference manual 2) 3) 4) 5) 11 Appendices Figure 5: PCB view from above on the left and from below 16

17 Figure 6: Different sensors (GPS, GyroZ, GyroXY, Accelero-meter) 17