Red Team. DARPA Grand Challenge Technical Paper. Revision: 6.1 Submitted for Public Release. April 8, 2004

Similar documents
Rover Systems Rover Systems 02/29/04

Super Squadron technical paper for. International Aerial Robotics Competition Team Reconnaissance. C. Aasish (M.

Control of Mobile Robots

Technical Paper Draft /13/2003. The Blue Team Anthony Levandowski

Technical Paper DARPA Grand Challenge Amendment/Update to October 2003 Technical Paper

Eurathlon Scenario Application Paper (SAP) Review Sheet

UNIVERSITÉ DE MONCTON FACULTÉ D INGÉNIERIE. Moncton, NB, Canada PROJECT BREAKPOINT 2015 IGVC DESIGN REPORT UNIVERSITÉ DE MONCTON ENGINEERING FACULTY

GCAT. University of Michigan-Dearborn

Detailed Design Review

Table of Contents. Abstract... Pg. (2) Project Description... Pg. (2) Design and Performance... Pg. (3) OOM Block Diagram Figure 1... Pg.

MAX PLATFORM FOR AUTONOMOUS BEHAVIORS

This is the technical paper required by the rules of the DARPA Grand Challenge Competition for 2004.

In 2003, A-Level Aerosystems (ZALA AERO) was founded by current company President Alexander Zakharov, since then he has led

Smart Control for Electric/Autonomous Vehicles

Odin s Journey. Development of Team Victor Tango s Autonomous Vehicle for the DARPA Urban Challenge. Jesse Hurdus. Dennis Hong. December 9th, 2007

UAV KF-1 helicopter. CopterCam UAV KF-1 helicopter specification

Autonomous Ground Vehicle Technologies Applied to the DARPA Grand Challenge

Problem Definition Review

Technical Overview of CajunBot (2005) Team CajunBot. DARPA Grand Challenge 2005

Our Approach to Automated Driving System Safety. February 2019

SAFERIDER Project FP SAFERIDER Andrea Borin November 5th, 2010 Final Event & Demonstration Leicester, UK

Technical Paper DARPA Grand Challenge 5 September, 2003

Jimi van der Woning. 30 November 2010

FLYING CAR NANODEGREE SYLLABUS

Automated Driving: Design and Verify Perception Systems

Technical Paper Team Mojavaton Car # 1 DARPA Grand Challenge 2005

Introduction Projects Basic Design Perception Motion Planning Mission Planning Behaviour Conclusion. Autonomous Vehicles

DARPA Grand Challenge 2005 Technical Paper. August 11, Race Date: October 8 th, Prepared by: Axion, LLC

Technical Paper for DARPA Grand Challenge

RIMRES: A project summary

3 DESIGN. 3.1 Chassis and Locomotion

DARPA Grand Challenge Virginia Tech Team Vehicle Name: Cliff Technical Paper

Wheeled Mobile Robots

Unmanned autonomous vehicles in air land and sea

Enabling Technologies for Autonomous Vehicles

Car Technologies Stanford and CMU

Development of an Autonomous Vehicle for High-speed Navigation and Obstacle Avoidance

AUTONOMOUS VEHICLES & HD MAP CREATION TEACHING A MACHINE HOW TO DRIVE ITSELF

AUTONOMOUS VEHICLES: PAST, PRESENT, FUTURE. CEM U. SARAYDAR Director, Electrical and Controls Systems Research Lab GM Global Research & Development

CS 188: Artificial Intelligence

Automated Driving - Object Perception at 120 KPH Chris Mansley

Adams-EDEM Co-simulation for Predicting Military Vehicle Mobility on Soft Soil

Preliminary Study of the Response of Forward Collision Warning Systems to Motorcycles

Electrical Dump Truck 980E-4

Vehicles at Volkswagen

Technical Paper DARPA Grand Challenge 2005

INTRODUCTION Team Composition Electrical System

High-accuracy Dead-reckoning System (HADRS) for Manned and Unmanned Ground Vehicles

Active Driver Assistance for Vehicle Lanekeeping

Technical Paper for Team Tormenta. DARPA Grand Challenge 2005

Holistic Range Prediction for Electric Vehicles

Project DIAS: Drexel s Integrated ATV System

IN SPRINTS TOWARDS AUTONOMOUS DRIVING. BMW GROUP TECHNOLOGY WORKSHOPS. December 2017

Virtual Testing of the Full Vehicle System

The Design of an Omnidirectional All-Terrain Rover Chassis

Crew integration & Automation Testbed and Robotic Follower Programs

BY HOEYCOMB AEROSPACE TECHNOLOGIES. HC-330 HYBRID-POWERED ALL- ELECTRICITY DRIVEN four-rotor UAV

University of Michigan s Work Toward Autonomous Cars

Team CIMAR DARPA Grand Challenge 2005 Sponsored by Smiths Aerospace

MEMS Sensors for automotive safety. Marc OSAJDA, NXP Semiconductors

Eurathlon Scenario Application Paper (SAP) Review Sheet

Development of a Multibody Systems Model for Investigation of the Effects of Hybrid Electric Vehicle Powertrains on Vehicle Dynamics.

THE FAST LANE FROM SILICON VALLEY TO MUNICH. UWE HIGGEN, HEAD OF BMW GROUP TECHNOLOGY OFFICE USA.

Formation Flying Experiments on the Orion-Emerald Mission. Introduction

The VisLab Intercontinental Autonomous Challenge: 13,000 km, 3 months, no driver

Deep Learning Will Make Truly Self-Driving Cars a Reality

RB-Mel-03. SCITOS G5 Mobile Platform Complete Package

Spectrum PowerCC DM Distribution Management The control system for distribution network operators

Prototyping Collision Avoidance for suas

Autonomous Haulage System for Mining Rationalization

Evaluation of Autonomous Ground Vehicle Skills

SIP-adus Field Operational Test

Remote Explorer (REx IV): An Autonomous Vessel for Data Acquisition and Dissemination

Cooperative Autonomous Driving and Interaction with Vulnerable Road Users

Fire Power Forum. DISTRIBUTION A: Approved for Public Release.

To put integrity before opportunity To be passionate and persistent To encourage individuals to rise to the occasion

UAV Enabled Measurement for Spatial Magnetic Field of Smart Rocks in Bridge Scour Monitoring

Y. Lemmens, T. Benoit, J. de Boer, T. Olbrechts LMS, A Siemens Business. Real-time Mechanism and System Simulation To Support Flight Simulators

Application of STPA to a Shift by Wire System (GM-MIT Research Project)

DD2-PRO+ Gps enabled Datalogger & display system mm (W) x 90mm (H) x 28mm (D) Datalogger Front 109mm (W) x 35mm (H) x 121mm (D) Datalogger Back

NASA Glenn Research Center Intelligent Power System Control Development for Deep Space Exploration

Oakland University Presents:

Development of the SciAutonics / Auburn Engineering Autonomous Car for the Urban Challenge. Prepared for: DARPA Urban Challenge

Items to specify: 4. Motor Speed Control. Head Unit. Radar. Steering Wheel Angle. ego vehicle speed control

KEEP TRACK OF WHAT MATTERS

The Heart of Passive Safety (Electronics) is Still Beating. Tom Borninski Autoliv Electronics America

RED RAVEN, THE LINKED-BOGIE PROTOTYPE. Ara Mekhtarian, Joseph Horvath, C.T. Lin. Department of Mechanical Engineering,

PROJECT IDEA SUBMISSION

Autonomous Quadrotor for the 2014 International Aerial Robotics Competition

A practical inertial navigation solution for continuous miner automation

Citi's 2016 Car of the Future Symposium

Co-Simulation of GT-Suite and CarMaker for Real Traffic and Race Track Simulations

WHITE PAPER Autonomous Driving A Bird s Eye View

Cat Electronic Technician 2017A v1.0 Product Status Report 7/10/2017 9:49 AM

Unmanned Surface Vessels - Opportunities and Technology

Technology for the Future of Vertical Lift

New impulses for sensing in automotive Dr. Richard Dixon

Financial Planning Association of Michigan 2018 Fall Symposium Autonomous Vehicles Presentation

What did we learn from Darpa Robotics Challenge

THE ULTIMATE DRONE SOLUTION

Transcription:

Red Team DARPA Grand Challenge Technical Paper Revision: 6.1 Submitted for Public Release April 8, 2004 Team Leader: William Red L. Whittaker Email address: red@ri.cmu.edu Mailing address: Carnegie Mellon University Robotics Institute 5000 Forbes Avenue Pittsburgh, PA 15213 Phone: (412) 268-6559 Fax: (412) 268-1338 Disclaimer: The views, opinions, and/or findings contained in this paper are those of the participating team and should not be interpreted as representing the official policies, either expressed or implied, of the Defense Advanced Research Projects Agency or the Department of Defense. DARPA and DoD cannot guarantee the accuracy or reliability of the information in this paper. The paper is published by DARPA as a service to those who seek additional technical information concerning the DARPA Grand Challenge of March 2004. Red Team Public Document ).

Table of Contents INTRODUCTION...1 1. SYSTEM DESCRIPTION...1 A. MOBILITY...1 B. POWER...3 C. PROCESSING...3 D. INTERNAL DATABASES...3 E. ENVIRONMENT SENSING...4 F. STATE SENSING...4 G. LOCALIZATION:...5 H. COMMUNICATIONS...5 J. NON-AUTONOMOUS CONTROL...5 2. SYSTEM PERFORMANCE...5 A. PREVIOUS TESTS...5 B. PLANNED TESTS...6 3. SAFETY AND ENVIRONMENTAL IMPACT...6 A. VEHICLE TOP SPEED...6 B. VEHICLE RANGE...6 C. SAFETY EQUIPMENT...6 D. E-STOPS...6 E. RADIATORS...7 F. ENVIRONMENTAL IMPACT...8 Introduction This paper describes the Red Team s technical approach for competing in the DARPA Grand Challenge and addresses 37 items outlined in http://www.darpa.mil/grandchallenge/technicalpaper.doc. These are divided among 3 sections: System Description, System Performance, and Safety & Environmental Impact. The paper is limited to 10 single-sided pages, exclusive of any cover pages, diagrams, or attachments. Four Attachments describe the vehicle design. The original GC questions are in Attachment B. GC Review comments are in Attachment C. 1. System Description a. Mobility The Red team s race vehicle, Sandstorm, is a M998 HMMWV vehicle modified for autonomous race driving (Figure 1). Red Team Public Document 1

Figure 1: Sandstorm March 2004 1. Ground contact is four wheels: 37" diameter, 12.5" width, 130" wheelbase, and 72" track width (see Figure 2). Widest part of vehicle is brush guard, 84. Suspension is coil over socket, Figure 3. 12.5 37 130 72 Figure 2: Footprint dimensions Figure 3: Suspension 2. 3. The vehicle is 4-wheel drive equipped, includes 4-wheel independent inboard disk brakes, and Ackermann steering for locomotion. Wheels are actuated through the stock drive train powered by a 6.2 liter turbo-charged diesel engine. Electric actuators control steering, throttle, braking, and gear shifting. Sensor stabilization is controlled by electric motors. Red Team Public Document 2

b. Power 1. The vehicle is powered by a 6.2 liter turbo-charged, diesel-fueled internal combustion engine. Vehicle electric bus is 24V alternator with dual series 12VDC lead-acid batteries. An auxiliary diesel generator provides payload power electrically filtered through lead-acid batteries. 2. Work is produced in two forms consisting of energy consumed as locomotive power and electrical consumption for vehicle computing, sensing and servo-mechanism systems. 3. At race start, the vehicle will hold sufficient diesel fuel to travel in excess of 300km. c. Processing 1. On-board vehicle computing systems defined as follows: Type Function Pentium III PC104 Stacks. Drive by wire control and supervisory, sensor vector pointing. Itanium 2 Based Server, 4 Path Planning Processors. Xeon Based Computer, LIDAR and RADAR data acquisition, terrain analyses, stereo Dual processors. processes, map data management and data transfer to path planner. 2. Perception perceived environmental interpretation is achieved by combining LIDAR, RADAR and stereovision range data. This data is then processed to synthesize a variety of maps used by vehicle logic to determine traversability. Obstacles are marked as high or infinite cost, while clear traverses are marked as low or no cost. Macro-route planning task is achieved by combining elevation topology, satellite imagery and ground truth data to form a single comprehensive global map. This map is then processed using a preplanning algorithm, generating a cost-map and route through areas of interest. Points on this route are then derived as goals for online planning. To facilitate synthesis of control arc and velocity, the algorithm considers multiple possible traversable arcs forward of vehicle position from the traversable space. Each possible arc is evaluated in terms of least cost to goal. The "best" path at any given interval is then communicated to the controller to command vehicle motion. d. Internal Databases 1. An off-board map database contains map features including sand, water, paved roads, unpaved roads, vegetation, rock, dry lake beds, out of bounds, and non-traversable terrain to the extent that they are known. This data comes from integrated USGS and BLM maps and is corrected relative to aerial imagery and road reconnaissance. During the two hour period prior to race, provided waypoints are used to extract relevant portions of this map database, Red Team Public Document 3

which is transferred to vehicle through a removable, wired connection. The distance from the external source to vehicle is within 150 feet. e. Environment Sensing 1. Perception sensors are listed in the table below with their primary functions, sensing horizon, and mode of operations. Sensor Long Range LIDAR line scanner, 1 each LIDAR line scanner, 3 each FMCW RADAR scanner Stereo video camera pair Active/Passive Primary Function Active Active Active Passive Terrain topology mapping, obstacle detection and characterization Obstacle detection and geometry characterization Obstacle detection Obstacle detection 2. Various sensor mounting configurations and functionality control scenarios as listed below: LIDAR line scanner 1. Vehicle roof mounted. 2. 3-axis stabilized platform mounted. 3. 3-axis pointing capability. 4. Custom interface controlled. LIDAR line scanner Mounted left and right, front of brush guard. 5. Mounted below radar and stabilized in pitch. 6. Controlled through RS422. RADAR scanner Mounted on the front of the E-box on a single axis stabilized platform operating in the pitch axis. 7. Shock/vibration isolated hard-mount. 8. Ethernet controlled. Stereovision camera pair Vehicle roof mounted. 9. 3-axis stabilized platform mounted. 10. 3-axis pointing capability. 11. Custom interface controlled. f. State Sensing 1. Vehicle state is sensed via optical encoders, potentiometers, rotational variable differential transformers (RVDT), current and voltage sensors. Vehicle state is reported by GPS (latitude, longitude, and altitude), vehicle Pose (roll, pitch, yaw), and vehicle velocity. Onboard software calculates vehicle s speed and acceleration. Red Team Public Document 4

2. Vehicle state sensors feed Vehicle Health, which produces a summary for Race Logic. The latter selects an appropriate race mode to guide control of the vehicle. g. Localization: 1. Vehicle localization is achieved via GPS/INS based Position & Orientation System, which globally estimates vehicle position and orientation with respect to map database. 2. The Position & Orientation System (POS) integrates last known vehicle position, encoded vehicle motion using a distance measurement indicator (DMI), and inertial navigation sensors (INS) to estimate and update vehicle position and orientation. 3. Areas outside the Grand Challenge Route boundaries are registered non-traversable on the internal map database. The vehicle path-planning algorithm will not direct the vehicle to traverse regions registered as non-traversable. 4. Applanix POS unit is utilized for inertial/gps/dmi instrumentation. The Applanix POS system is a strapdown inertial navigation platform, featuring high-bandwidth, low-latency, Kalman filtering, GPS with azimuth measurement, distance measurement indicator (DMI). Applanix platform calculates vehicle position and orientation data with excellent short-term accuracy and minimum long-term drift. Applanix POS platform provides dynamically accurate, high-bandwidth full-kinematic vehicle state measurements, and provides motion compensation information to other sensor systems onboard the Grand Challenge vehicle. h. Communications 1. Prior to the race, pit crew to vehicle communications will utilize 802.11b/g wireless network and CAT5 Ethernet cable for debugging and setup. 2. At the race starting line, the 802.11b/g wireless communication network and the CTA5 Ethernet cable shall be removed prior to race start. Final instructions to the vehicle are communicated via CAT5 Ethernet cable. 3. During the race, vehicle will only receive DARPA E-Stop and Pause, GPS, and commercially available DGPS corrections via satellite. i. Non-autonomous control The vehicle will remain human drivable. The vehicle will also be controllable through a wireless Ethernet connection. The wireless Ethernet Bridge shall be removed prior to the race. 2. System Performance a. Previous Tests Extensive testing and evaluation was conducted to evolve vehicle sensing and autonomous steering capability. Test and evaluation demonstrated object and obstacle detection as well as proof-of-concept associated with perception sensing capabilities. Initial vehicle implemented integrated differential GPS, three LIDAR scanners and a video camera. Combining GPS, IMU Red Team Public Document 5

and LIDAR data facilitated building of 3D point-clouds showing obstacles, vegetation, and roads. Post-processing synthesized goodness-maps and facilitated run-simulations in addition to physical testing. Both wireless and mechanical E-stop systems were implemented and tested. Although the wireless E-stop was not specific unit provided by GC officials, proof of concept was successful. E-box cooling system was designed and implemented, based on analyzed and measured thermal characteristic data. E-box shock isolation system was designed and implemented, based on analyzed and measured dynamic inertial data. DARPA GC E-stop system was also implemented and verified functional. b. Planned Tests Incremental testing regime will continue as program develops and moves towards higher navigational speeds, more complex real-time processing, and increased sensing capability. Vehicle testing programs will include component, subsystem, speed, and desert local. 3. Safety and Environmental Impact a. Vehicle Top Speed The vehicle has a top speed adequate to complete the prescribed course within the allotted duration. b. Vehicle Range The vehicle has a range in excess of prescribed course. c. Safety Equipment The safety equipment on-board the Challenge Vehicle is listed below. 1. Fuel containment: Fuel line cutoff solenoid connected to the disable E-Stop. 2. Fire suppression: Fuel valve cutoff and multiple fire extinguishers mounted external to the vehicle. 3. The fire extinguishers shall be removed prior to registering for QID. 4. Audio and visual warning devices: Klaxon, warning strobe, and brake lights in compliance with DARPA requirements. d. E-Stops 1. While the normal e-stop signal is active, the vehicle controller will remain in an inhibited state, ignoring all higher-level control inputs. In this inhibited state, all electronics and actuators remain powered, and the engine continues to run, but the vehicle will not move. Upon release of the normal E-Stop signal, the vehicle controller returns to its normal operating state, accepting control input from higher level planning modules. The disable e- stop causes a normally open relay to actuate. This causes the vehicle engine to shutdown and the service brake is fully engaged, effecting the vehicle to an abrupt halt. Field personnel Red Team Public Document 6

will unlock the brakes by throwing a clearly labeled pneumatic lever switch to the OVERRIDE position that will release the parking brake, rotate a clearly labeled selector witch to the DISABLED position to cut power to control motor, and pull a clearly labeled cord that moves the service brake actuator off the brake pedal. The auxiliary generator will continue to run but can in no way cause the vehicle to move. 2. Manual e-stop switches are of industrial grade. E-stop switches are placed on left and right side amidships and placed on each corner of the vehicle for total of six (6) units. The six manual E-stop switches are wired in series and inclusive to the wireless disable E-stop signal line. This will cause the manual E-stop switches to behave in exactly the same way as the wireless disable E-stop signal. 3. The vehicle is placed in neutral by shifting the transfer case lever to the neutral position. However, due to the automatic transmission the vehicle can be moved without placing the transmission or transfer case in neutral. In addition, field personnel will be required to complete four steps: a. Throw a clearly labeled pneumatic lever switch to the OVERRIDE position; b. Move the "Brake/Throttle Actuator Switch" to the disable position c. Pull a clearly labeled cord to disable the safety brake function. d. Mount the 11" quick release removable steering wheel. e. Radiators 1. The devices on the Challenge Vehicle that actively radiate EM energy are listed below. Long Range LIDAR line scanner Short Range LIDAR line scanner RADAR scanner Vehicle and DARPA operational lights 3. The devices on the vehicle that can be considered hazardous to human eye or ear safety are listed below with their OSHA classification. Long Range LIDAR line scanner OSHA Class 1 eye safe. Short Range LIDAR line scanner OSHA Class 1 eye safe. RADAR scanner ETSI Class 1 radiator. 4. The following safety measures are important regarding the radiators. Avoid standing in close proximity to RADAR platform. Avoid looking directly at LIDAR scanner. Red Team Public Document 7

Avoid looking at rotating beacon and wear audio protection during vehicle operations f. Environmental Impact 1. When a tire flat occurs, the run flat inserts are non-metal and keep the metal rims from contacting the paved roadway. 2. The maximum ground pressure is 32psi while tire are inflated. 3. Engine crankcase, engine cooling system, transmission and transfer case, differential gear housings, power steering and hydraulic brake system, and fuel tank contain various fluids, which pose a potential threat to the environment. Red Team Public Document 8