Student Projects
Contact Information
University and Department: Worcester Polytechnic Institute, Robotics Engineering Program
Team Name: Prometheus
Primary LabView Developer: Daniel Sacco
Other Team Members: Justin Barrett, Rob Fitzpatrick, Chris Gamache, Ricardo Madera, Adam Panzica, Benjamin Roy, Viktoras Truchanovicius, Bohua Wang
Primary Faculty Advisor: Professor Taskin Padir
Primary Email Address: dsacco@wpi.edu
Primary Telephone Number (include area and country code): 001 973 769 3248
Team Website: www.igvc-wpi.org
Project Information
Project Title: Design and Realization of an Intelligent Unmanned Ground Vehicle
List all parts you used to design and complete your project:
Software:
- LabView 2009
- Netbeans 6.8
- Eclipse Galileo
- CUDA Development Kit
- Windows XP
- Windows 7
- Ubuntu 9.10
Hardware:
- NI-9074 CompactRIO
- NI-9472 Sourcing Digital Output Module
- NI-9201 Analog Input Module
- NI-9870 RS-232 Serial Interface Module
- NI-9403 Digital Input/Output Module
- EVGA E758-TR Motherboard
- NVIDIA Tesla C1060
- SICK LMS291-S05 LIDAR
- SOKKIA Axis 3 DGPS Receiver
- Point Grey FL2G-13S2M/C Cameras (2)
- PNI V2xe 2-Axis Compass
- NPC-T64 Drive Motors (2)
- Luminary Micro Black Jaguar Motor Controllers (3)
- US Digital Quadrature Encoders
- Optima YellowTop Batteries (2)
Describe the challenge your project is trying to solve.
Robotics and intelligent systems are relatively new areas of technology that have been gaining interest from the military. Soldiers put themselves in danger when transporting materials, performing reconnaissance missions and performing search missions. In these instances, an autonomous intelligent ground vehicle would provide the ability to replace people with robots and ultimately save lives by keeping soldiers out of dangerous situations.
Advancement of the technologies involved in intelligent ground vehicles provides great benefits to society too. Currently, motor vehicles are dangerous and prone to accidents due to human error. In 2008 alone, there were 34,000 fatal car accidents (FARS. Fatal Accident Reporting System, 2008). Navigation, mapping and object detection technologies have the potential to increase vehicle safety and prevent such accidents. Many automotive companies are incorporating intelligent systems into their newest models for a safer and more comfortable driving experience. Driver assistance technology which continuously evaluates the surroundings of the vehicle and provides information to the driver can also take control of the vehicle if needed. Smart cruise control, collision warning and airbag systems can potentially save thousands of lives daily and increase driving pleasure.
Our project focuses on designing and building an intelligent unmanned ground vehicle (UGV) which will become WPI’s first entry to the Intelligent Ground Vehicle Competition (IGVC) in June 2010. IGVC challenges students to build and program a fully autonomous UGV that can locate and avoid obstacles, stay within the boundaries of a lane, navigate to GPS waypoints and implement a communications system using the Joint Architecture for Unmanned Systems (JAUS) protocol all while carrying a payload.
Describe how you addressed the challenge through your project.
WPI’s intelligent UGV, Prometheus has a custom welded-aluminum chassis with two rear differential drive wheels and a steered front wheel. The vehicle power train consists of sealed lead acid batteries, bi-directional DC motor and chain sprocket sets. For obstacle detection, the vehicle utilizes a LIDAR sensor. For navigation to GPS waypoints a differential GPS unit is used along with a 2 axis compass and information obtained from wheel encoders. Line and pothole detection is accomplished using stereovision cameras and computational power is obtained from a National Instruments cRIO and an ATX style computer. The vehicle design can be seen in Figure 1.
Figure 1: Vehicle Design and Components
The hardware used for vehicle control employs a distributed system approach using the ATX style computer and the cRIO controller. Using this approach, the responsibilities of the systems are divided to allow for parallel development and parallel execution of tasks which increases development and system performance. The computer is responsible for performing stereo image processing implemented on NVIDIA's GPU. The image processing algorithms include segmentation using neural networks, rectification and pixel disparity calculation for line and obstacle distance detection. The computer is also responsible for navigation planning which is accomplished by creating an obstacle probability map and path planning is implemented based on the Driving with Tentacles approach. The NI cRIO is responsible for sensor fusion which incorporates data from a LIDAR, differential GPS receiver, wheel encoders, and a compass. The cRIO is also responsible for executing the path that is determined by the computer. The distributed system communicates over Ethernet so that data the computer can constantly be updated with the newest sensor information and the cRIO with the newest path information. A diagram of the distributed system can be seen in Figure 2.
Figure 2: Distributed System Diagram
The software running on the cRIO is divided into two main sections, the CPU program and the FPGA configuration. Using LabView, programs were successfully created on for both of these targets to
- Receive and convert LIDAR object data
- Receive and convert GPS data
- Read quadrature encoders
- Read heading information from 2 axis compass
- Establish PID motor control
- Communicate over ethernet using UDP and custom packet structure
- Execute a given driving path
- Switch between vehicle autonomous and manual modes
- Use a Kalman Filter for data fusion to determine accurate vehicle local position
Final Vehicle Image:
Figure 3: Vehicle Outside
Video clip of vehicle driving through obstacle course:
