An Autonomous Ground Vehicle

Download PDF
5 downloads 0 Views 527KB Size Report
Comment
The main technological challenge in the development of vehicle is to build a highly .... Autonomous Vehicle for DARPA's Grand Challenge. Robert Grabowski ...
VECTOR2703

An Autonomous Ground Vehicle Tushar Chugh, Ms. Meenakshi Gupta Student, Asst. Prof., MRCE, Faridabad [email protected],[email protected] ABSTRACT This document presents an approach towards designing an autonomous ground vehicle Vector2703 that can be used by military for low intensity conflict. It the team’s approach to building its robotic vehicle, the vehicle and components that let the vehicle see and act, and the computer software that made the vehicle autonomous. It presents how the vehicle is going to be designed and will work in various fields of the arena we have decided.

in the competition with no hindrance. Vector2703 has a rear wheel drive and has centrally mounted 110 CC

Keywords: Vector 2703, Sensors, Software architecture.

I. INTRODUCTION Team AutoMatrix is working on designing Autonomous Ground Vehicle Vector2703 for military purposes that can be used for the lowintensity conflicts. The primary goal of this vehicle is to navigate autonomously through the path given decided by us. The specifications of this robotic vehicle are kept in mind keeping in view the rules and regulations we have decided. Vector2703 is based on the self made mechanical design outfitted with a computing platform and a suite of sensors and actuators provided for the autonomous navigation. The main technological challenge in the development of vehicle is to build a highly reliable system, capable of driving at relatively high speeds through diverse and unstructured off-road environments, and to do all this with high precision. We are working on the method to developed long-range terrain perception, real-time collision avoidance, and stable vehicle control on slippery and rugged terrain. This system is achieved by integrating various techniques used in various branches of engineering. II.VEHICLE Vector2703 has been designed and carefully simulated to make sure that it traverses the path given

4-stroke petrol engine with CVT type transmission. Cylinder used is a naturally air-cooled single cylinder. Body of the vehicle is constructed using the seam less pipes with the dimensions 20id * 22od. For proper stability and keeping in view the centre of mass the weight is equally distributed on the wheels. The suspension employed in the vehicle is 4-wheel independent suspension that provides unique technical advancement for the all terrain purposes. It ensures a good tractive force between wheel and ground and better stability over uneven tracks resulting elimination of chances of skidding or toppling. The steering control is through spur gear set actuated by servo motor. Throttle and Brakes are electronically actuated by servo motors. The wheel base is 32 inches and ground clearance is 10.5 inches. The wheels have diameter of 15 inches to ensure easy climbing while on the course of stairs and corrugations. Swing arms are provided to evade any chances of toppling. Vector2703 is outfitted with 3 CCD cameras, 3 ultrasonic sensors and a global positioning system with compass. One CCD camera is mounted in the centre at the height 0.60m and the other two sensors are placed at left and right of the central camera at the

same height of 0.48m. Arrangement of the ultrasonic sensors is also depicted in the picture. One sensor is facing downwards at the specified angle while the other two sensors are placed at proximity to the wheels. The front portion of the vehicle has a battery for running the electrical circuits and servo motors (Battery used in Cars: 12V, 42Ah). We also have a provision for payload at this end only. In the rear end we have the housing for one computer and other electronic circuits which are used as the processing units.

A. TRANSMISSION Transmission is CUT type accompanied with chain drive for torque enhancement. Torque enhancement through the system is 25 times.

ULTRA SONIC SENSORS(ULS) Ultrasonic signals are like audible sound waves, except the frequencies are much higher. Ultrasonic transducers have piezoelectric crystals which resonate to a desired frequency and convert electric energy into acoustic energy and vice versa. Variables which can affect the operation of an ultrasonic sensor include: target surface angle, reflective surface roughness, and change in temperature or humidity. The targets can have any kind of reflective form and even round objects are an acceptable target. Advantages of Ultrasonic Sensors 1. Discrete distances to moving objects can be detected and measured. 2. Less affected by target materials and surfaces. 3. Not affected by color. 4. Solid state – virtually unlimited maintenance- free life. 5. Small objects can be detected over longer distances. 6. Resistance to external disturbances such as vibration, infrared radiation, ambient noise and EMI radiation. Application and Interfacing Vector2703 is optimized with three ultrasonic sensors for the obstacle detection.

Figure: Drive Train III.SENSORS

The positioning of the sensors is shown in the pictorial view of the vehicle. First sensor (ULS1) is placed at a height of 0.48m above the ground with an angle of 4.716 degrees with the horizontal as shown in the figure. Second (ULN2) and third (ULN3) sensors are mounted at extreme left and right side of the vehicle at the height of 0.30m. The arrangement makes sure to detect and differentiate between static obstacles, stairs, 15 degree positive slope and corrugation. Detection of obstacle by the middle sensor is depicted by the figure below:

Figure: Ultrasonic Sensor Positioned at the Centre B. VISION SENSORS- CCD CAMERA C. NAVIGATION SENSORS - GPS AND COMPASS

Vector2703 uses CCD Camera’s as the vision sensors. CCD Camera’s convert light into electric charges and process it into electrical signals. In a CCD sensor, every pixel is transferred through a very limited number of output nodes to convert into voltage, buffered, and the output’s uniformity (a key factor for image quality) is high. CCD offers excellent imaging performance benchmarks in photography related to scientific research and other important areas. Vision Functioning Vector2703 operated using three CCD Camera’s which functions collectively to detect the path by color sensing deployed on the sides of the path as well as these detect the obstacles. The positioning of the 3 camera’s are shown in the picture. The function of the middle camera (CAM1) is to give an idea of the topography as well as some details about the obstacle. Function of CAM2 and CAM3 is to detect the yellow and black strips provided along the path. These are inclined below w.r.t horizontal so as get the image of that require strip enough to follow the path. The outputs of CAM1 plus the output of three ultrasonic sensors provides complete details about the topography and obstacles coming in front.

Positioning for the Vector2703 is accomplished by one GPS unit (Global Positioning System) and an electrical compass. The GPS unit uses GLOBALSAT ET-314 as the GPS Engine Board. This is a compact, high performance, and low power consumption (voltage required is 3.3 V) GPS engine board. It uses SiRF Star III chipset which can track up to 20 satellites at a time and perform TTFF (Time to first fix) in weak signal environments. It supports NMEA 0183 and SiRF binary protocols. Here we will be using NMEA 0183 v2.2 protocol (GGA, GSA, GSV, RMC, GLL, VTG). Serial port is used for communication with a baud rate of 9600. As its accuracy is of 10 meters in the normal mode so this is to be operated in the differential mode (DGPS). By the DGPS we get the accuracy as 3 meters (WAAS mode). Therefore we can choose one of the protocols let’s say GLL to determine various parameters (latitude, longitude, UTC time, altitude and velocity). As the GPS tell about the location (lat, long) but not the angle of the vehicle therefore we have to use some other device for this reason. Therefore, we use an Electric compass to find out the actual direction of the vehicle. This gives the output in the electric form which can be processes easily. Now by both the devices we can know the overall position of the vehicle. Next we also know the position of the next or target waypoint.

If we draw a vector from initial to the final position then we can actually know where we have to proceed. If the vehicle is running from the input then it is first aligned in parallel to the vector then it starts navigating towards the next waypoint or the target. Waypoints are stored in the memory and the processing of the input is done in the software and the corresponding output is generated for the actuators. We can create a virtual map using the waypoints provided and the vehicle has to travel according to that virtual map. The navigation using GPS is given higher priority if the side camera’s are not sensing any yellow-black strip and also the obstacles can be located using the waypoints.

GPS MODULE IV SOFTWARE ARCHITECTURE Software Architecture is one of the main areas which are to be emphasized. So after integrating the above mechanical and electronic modules we have to treat autonomous navigation as software problem. The hardware diagram explains how various input devices are connecting to the processing unit which processes the information and finally it gives the output to the actuators to control steering, throttle and brakes. Figure shows that inputs devices are CCD Camera, Ultrasonic Sensors, Electric compass and GPS which gives the output to the computer in which we have the software application which process the input and generates a set of commands to control the actuators. These commands are sent to microcontroller which controls the servo motors. Therefore, actuators are controlled by the output of the microcontroller. The speed of each wheel is also calculated by tachometer using microcontroller and this measurement is sent as an input to the computer

so that it acts as a feedback and we can monitor the speed. Vector2703 also has the provision for E-stop (Emergency Stop) controlled by RF remote and is connected directly to the actuators so that if there is an emergency when the vehicle gets out of control then by this control we can stop the vehicle.

A. DATA FLOW AND ALGORITHM The whole software problem has been simplified and based on that we have made a flow control diagram. This flow diagram fully explains how the data/signal flows in the from the sensor interface to the actuators. We have categorized the data flow into four sections as shown in the diagram. SENSOR INTERFACE: This section is the primary section which gives us the electrical signal about the information of the surroundings (ultrasonic sensors, CCD CAM, GPS, Compass, Tachometer, angle of steering and brakes ) PERCEPTION: The signals from the sensors re sent to the computer/processor for processing the input. This communication is done by the ports like CCD CAMERA- USB GPS- Serial port. The software is designed in MATLAB which is deployed as standalone application. Here the signals from CAM1, CAM2, and CAM3 are processed using image processing toolbox(Vision Mapper) and interpretation about path (side strips, Obstacles) is made in this loop. In the second loop the signal from the ultrasonic sensors is sent to the ULS mapper (User created) and gives the information about the obstacles. In the third loop position of the vehicle is estimated by the inputs of GPS and Compass using the vector mapping.

PLANNING AND CONTROL: The output of the perception section is fed to this section. This means its input is the output of the three loops stated. Path planning in this section is done by resolving the priorities of the inputs of the three loops. Highest priority is given to the obstacle voidance, next is the lane following and finally is no obstacle and colored strips then the output is controlled by Position Estimating Mapper loop. Then the output of this is given to the actuators and steering, brakes, throttle are controlled by the servo motors (by interfacing with 8051 microcontroller). USER INTERFACE: Although the vehicle is autonomous we should not have any user interface but for the safety requirements we should have an emergency switch which is controlled by RF Remote. There are also mechanical switches which can switch transmission between auto and self modes.

REFERENCES 1. MITRE Meteor: An Off-Road Autonomous Vehicle for DARPA’s Grand Challenge Robert Grabowski, Richard Weatherly, Robert Bolling,David Seidel, Michael Shadid, and Ann Jones. http://citeseerx.ist.psu.edu/viewdoc/download? doi=10.1.1.133.8823&rep=rep1&type=pdf 2. Stanley: The Robot that Won the DARPA Grand Challenge Sebastian Thrun, Mike Montemerlo,Hendrik Dahlkamp, David Stavens,Andrei Aron, James Diebel, Philip Fong,John Gale, Morgan Halpenny,Gabriel Hoffmann, Kenny Lau, Celia Oakley, Mark Palatucci, Vaughan Pratt,and Pascal Stang http://marcmbrs.ucsc.edu/documents/thrun.stanley05. pdf 3. MOBILE MAPPING AND AUTONOMOUS VEHICLE NAVIGATION C. K. Totha, E. Paska baOSU, Center for Mapping, 1216 Kinnear Road, Columbus, OH 43212-1154, USA – [email protected] bOSU, Department of Civil and Environmental Engineering and Geodetic Science, Columbus, USA http://www.isprs.org/proceedings/XXXVI/part1/Pape rs/T08-36.pdf

Figure: Flowchart for Vector2703 software model

WORK DONE SO FAR I have worked on the navigation part with GPS system and have integrated the gps receiver with micro-controller. Presently the work on the designing part has been started and first the robotic model based on track drive is under the process.