New Jersey Autonomous Vehicle Brandon Schiff Jason Scott

New Jersey Autonomous Vehicle Brandon Schiff Jason Scott Jared Milburn

Abstract � Comprised of one mechanical and two computer engineers � Construct vehicle to navigate through an obstacle course by GPS Waypoints � Improving � Compete previous model in the 22 nd Annual IGVC

Table of Contents � IGVC � Frame Design � Electronics � Alogrithm � Future � Budget Design Plans

Intelligent Ground Vehicle Competition (IGVC) � The 22 nd Annual Intelligent Ground Vehicle Competition � Oakland University in Rochester, Michigan � June 6 – June 9, 2014 � Ground Vehicle � Autonomous � Qualification � Basic and Advance Courses

IGVC Rules and Regulations � Size ◦ Length – 3 ft-7 ft ◦ Width – 2 ft-4 ft ◦ Height – Under 6 ft � Speed ◦ Average – 1 mph ◦ Minimum – 1 mph ◦ Maximum – 10 mph � Propulsion � Emergency Stop ◦ Wireless ◦ Mechanical � Safety Light � Payload ◦ 18” x 8” ◦ 20 Pounds

IGVC Courses � Grass with Dashed Lines � Natural and Manmade Objects � Waypoints � Colored � Fencing Flags

IGVC Courses

Frame Design � Previous � Stress Frame Analysis � Compliance � Material with IGVC Rules Used

Analysis of previous team’s frame: Left and right: deformation caused by load and laser range finder

IGVC Spec. Dimensio n Length Width Height TCNJ Autonomous Vehicle 3’ 2. 5” 2’ 2” 3’ 6” • Still the best design iteration • Functional design • Cons have simple solutions • Allows focus to be shifted to ensuring vehicle is fully operational • Blue Loctite used to lock bolts in place IGVC Specifications 3’ - 7’ 2’ – 5’ Max: 6’

Material 6105 T 5 Aluminum Fractional T-slotted bars Product Number: 1010 Cross Sectional Dim. : 1. 00” x 1. 00” E = 10, 000 ksi ν = 0. 33 Reasoning: • Budget Friendly • Lightweight • Machinable • Modular

Electronics Overview � Allows the vehicle to be aware of it’s environment and location � Powered by two separate on-board batteries or laptop. � Laptop used for data processing of electrical components

Drive Train Diagram Manual E-Stop Button 12 Volt Battery Relay Wireless E-Stop Button Motor Controller Motor Optical Encoder Microcontroller

Drive Train Propulsion � Four Wheels, Two Wheel Drive � NPC-42150 Motors ◦ ◦ DC Motors Torque - 100 Psi 93 Rpm Previous Years ◦ ◦ Model – Sabertooth 2 x 25 V 2 Controls both motors Controlled through serial ports Previous Years � Motor Controller

Feedback System � Measure Wheel Speed � Optical Encoder ◦ Attached to gear shaped Disk ◦ LED Light ◦ Voltage Pulses Voltage No Tooth

Global Positioning System (GPS) � Used to navigate vehicle to given GPS location � Data sent via serial connection to Arduino port � Used in accordance with magnetometer

Digital Compass � Reads � Digital current vehicle orientation as opposed to analog compass � Accompanies � Arduino GPS system serial connection and power

Webcam � Used to feed real time images of the course to our laptop � Primarily focused on line detection as opposed to object detection � Filters out unnecessary visual information through applying masks and focuses only on discovering white lines � Recognition of white lines fed into path planning algorithm

Laser Range Finder � Short range laser used for object detection � Properties � Data sent via RS 232 -to-USB connection with laptop � Output

Power Systems � Laser Range Finder/GPS operating on two 12 V batteries � Compass/Webcam/Warning Light/Motors and Motor Controller running on 12 V � Sensors and vehicle operations communicates with Arduino Mega � Software-processing laptop sends and receives data with Arduino

Microcontroller � Arduino Mega � Outputs 3. 3 V and <50 m. A � Powered and communicates with laptop via USB � Arduino IDE

Caution Light Laser Range Finder Camera Laser Range Finder Software GPS Microcontroller Compass Communication Hub Arduino Software RC and RC Controller D* Lite Motors - Software Camera Software C++ (Eclipse IDE) - Hardware Components - Arduino - Laptop

Algorithm Design � Algorithms for the autonomous vehicle need to be robust and simple � Navigation and Path Planning algorithms are required for optimal performance � Navigation algorithm relies on utilizing the capabilities of the GPS and Compass while the Path Planning algorithm relies on the webcam and laser range finder

Navigation � Determines the vehicle’s current position, maintains a list of waypoints, and keeps track of the vehicle’s progress � GPS must accurately determine and report the vehicle’s latitude and longitude � Compass heading must give the vehicle’s current

Path Planning � Going � D* to use D* Lite path planning is an assumption based algorithm useful for when a robot needs to navigate to a given goal in unknown terrain � D* Lite works with the same functionality as D*, but it is simpler to understand easier to execute

Path Planning

Software Used � Previously ◦ ◦ Matlab Microsoft Visual Studio Open CSV Arduino IDE � Now ◦ Eclipse C++ Language IDE ◦ AVR-GCC Compiler ◦ AVRdude

Software Goals � Reproduce � Testing � Write all MATLAB code in C++ of C++ code path planning and navigation algorithms � Final program formulated using Microsoft Visual Studio and Open. CV

Future Plans � Frame covering � Full electrical system finalized, connected, and run simutaneously � RC controller configuration and testing � New coding, testing and debugging

NJAV Future Plans � Spring ◦ Finalize Frame and Drive Train ◦ Path Planning Components Working in Sequence ◦ Debugging and Testing � Summer ◦ Final Testing and Preparation for IGVC

Budget Total Price Mechanical $100. 00 Electrical $122. 90 Travel $1136. 00 Total $1358. 90

Acknowledgements � Dr. Jennifer Wang � Dr. Orlando J. Hernandez ◦ Advisor – Professor of Mechanical Engineering – The College of New Jersey ◦ Advisor – Professor of Electrical and Computer Engineering – The College of New Jersey � Mr. Joseph Zanetti � Dr. Steven Schreiner ◦ Professional Services Specialist – School of Engineering – The College of New Jersey ◦ Dean of the School of Engineering – The College of New Jersey

Questions? � New Jersey Autonomous Vehicle ◦ Jason Scott ◦ Jared Milburn ◦ Jonathan Sayre