Autonomous Surface Navigation Platform

1
Autonomous Surface Navigation Platform

• Michael Baxter
• Angel Berrocal
• Brandon Groff

2
Project Overview

• General platform capable of autonomous navigation
  and adaptable to land and sea surface
  environments.
• Apply sensor technology for a basic navigation
  software foundation on which to build more
  complex autonomous algorithms for homeland
  security. Primary client is multiple branches of
  the US Military.
•  
• Estimate development costs at 41,000 initial
  cost of a prototype system, including
  instrumentation and wiring, should be less than
  10,000.

3
Technical Objectives

• Basic Autonomous Navigation
• Collision Avoidance
• Path Planning and Following
• Propulsion System Control
•  
• Recognition of Objective
• Object Color Shape Recognition
•  
• Built Using LabVIEW CompactRIO Platforms
• Readily available
• LabVIEW Vision Module is easier to use than other
  solutions

4
Demonstration of Objectives
5
Navigation System Design
6
Actual Design Implementation
7
Overall Program Flow
8
Path Planning

• Requirements
• Enable vehicle to drive toward target object by
  determining direction in which it should proceed
• Requires a devices to determine the current
  position and angle of the robot.
• GPS Device - Determination of absolute position
• Inertial Motion Unit (IMU) - Information about
  robot yaw
• Software Interface Requirements
• LabVIEW 2009
• CompactRIO Serial Module for GPS Interface
• Caveats
• GPS Module has accuracy of 3 meters or less
  using Wide Area Augmentation System (WAAS)

9
Path Planning Algorithm
10
Path Following

• Requirements
• Must allow robot to follow a planned path
• May initiate object avoidance while following
  path
• Needs sensors to track current heading, and watch
  for potential obstacles
• IMU - Track heading can be reused from Path
  Planning
• Vex Sonar - obstacle detection avoidance
• Trade offs
• Shaft Encoder - Less expensive, require more
  programming, propulsion-system dependent

11
Path Following Algorithm
12
Collision Avoidance

• Requirements
• Autonomous navigation without collisions which
  may stop, damage or otherwise hinder the
  vehicle's movement
• Requires sensor for detection of distances to
  objects
• Laser Range Finders - very expensive
• Vex Ultrasonic Sonar - very cheap and available
• Software Requirements
• LabVIEW 2009
• LabVIEW FPGA Module (sonar interface)
• Caveats
• Cheaper sensors require better software

13
Collision Avoidance Algorithm
14
Object Search Recognition

• Requirements
• 360-degree view to scan for possible object color
  and shape recognition
• One camera facing forward and one facing backward
  for improved search time
• Requires two cameras that are compatible
  with micro-controller CompactRIO
• Axis M1011 Ethernet cameras can be used
  simultaneously with CompactRIO
• Wireless-G Linksys Router 2.4 GHz 
• Software Requirements
• LabVIEW 2009 LabVIEW Vision Development Module
• Caveats
• Better cameras with higher resolution are costly
• Low resolution requires better image processing 

15
Object Search
16
Object Recognition Algorithm

• LabVIEW Vision Module allows high-level
  development of vision algorithms in a
  block-diagram fashion

• LabVIEW Vision Assistant Block Diagram

17
Propulsion System Control

•  Requirements
• Must be able to control Thrusters for marine
  vehicles as well as Vex motors for land
  demonstration
• Both use 1-2 ms Pulse-Width Modulation signals
• Requires hardware interface between LabVIEW and
  propulsion system
• CompactRIO Digital I/O Module
• Software Requirements
• LabVIEW 2009
• LabVIEW FPGA Module

18
Schedule of Future Work 
19
Design Issues

• GPS Device
• Payment issues purchasing first GPS CompactRIO
  module
• Received a pledge for a donated GPS CompactRIO
  module from Marine Innovations, however their
  driver appeared to be for an old version of
  LabVIEW
• Package from Marine Innovations showed up empty
• Just purchased a Garmin stand-alone unit with
  serial interface
• Vex sonar inaccuracies
• Sonars are relatively cheap and often provide
  inaccurate readings - requires averaging and
  filtering of data

20
Summary of current status of project

• Vehicle Status
• Test platform assembled
• Motors, Cameras, Router, IMU, Sonars, and cRIO
  mounted
• Garmin GPS has not arrived, driver available from
  National Instruments
• Power Plan designed - 24V 7.2V batteries with 2
  DC-DC converters
• Algorithm Status
• Able to gather usable data from sonars, cameras,
  and IMU 
• Object recognition algorithm successfully
  differentiates between similar objects indoors
  and determines distance to object
• Successful propulsion system control with cRIO

21
LabVIEW Test Program Example