Research in Autonomous Mobile Robots: Research in Autonomous Mobile Robots Presented by: Anantharaman, Divya Hogan, Rose Mahtani, Amrit Tiwari, Manjari Scott, Kellum Wang, Ling Project Guide Dr. Peggy Israel Doerschuk Supported by grants from Exxon Mobil & The State of Texas.


Problem Statement: Problem Statement This project is aimed at building a robot to integrate reactive behavior and deliberative planning. The robot’s task is to locate a nest by distinguishing the color of the nest and its surrounding . It must then detect an object, pick it up and randomly place the object in the empty nest while avoiding obstacles in its path. Once the object is placed in the nest the robot must remember that the nest is occupied.


Outline: Outline Problem Analysis Hardware Basics Software Architecture Basics Find Object Find Nest Get Object Deposit Go to Nest Avoid Obstacle Conclusion


Problem Analysis: Problem Analysis


Hardware Used: Hardware Used Handy Board Expansion Board LEGO Sensors CMU Cam 1 Rotation Sensor Servo Motors Light Sensor DC Motors Proximity Sensor


Handy Board: Handy Board Motorola 68 HC11 CPU 8 bit processor 32K of external battery backed up RAM 16 I/P for digital and Analog sensors 4 O/P for DC motors LCD screen: 16-char, 2-line liquid crystal display


Expansion Board: Expansion Board Allows use of 6 servo-motor controls Additional sensor inputs Digital outputs, and a pass-through connector for the LCD panel


Lego: Lego


Outline: Outline Problem Analysis Hardware Basics Software Architecture Basics Find Object Find Nest Get Object Deposit Go to Nest Avoid Obstacle Conclusion


Hybrid Deliberative/Reactive Control: Hybrid Deliberative/Reactive Control Under this Paradigm, the robot 1) Plans the task – Local Planning Layer. - how to decompose a task into subtasks - What are the suitable behaviors to accomplish each subtask. 2) Executes the task - Behavior based Layer. -Incorporates various patterns of behavior -Performs jobs requested by LPL through procedural POBs -Handles emergency situations through reactive POBs World Interface 1) Sensors & Motors


Pattern Of Behaviors: Pattern Of Behaviors


State Diagrams: State Diagrams


Outline: Outline Problem Analysis Hardware Basics Software Architecture Basics ROBOT BEHAVIOURS Find Object Find Nest Get Object Deposit Go to Nest Avoid Obstacle Conclusion


CMU Cam 1: CMU Cam 1 Major functionalities: 17 frames/ second Track the position and size of a colorful or bright object Measure the RGB statistics of an image region Low Cost, Low power Can be purchased online from Acroname, Seattle Robotics and Lextronic. Carnegie Mellon University Camera


CMU CAM GUI Screen shot of the software: CMU CAM GUI Screen shot of the software


Servo Motors: Servo Motors Components DC motor Gear Reduction Unit Shaft Position Sensor Electronic circuit that controls the motor’s operation. Feedback control unit


Find Object State Diagram: Find Object State Diagram


ALGORITHM - Find Object: ALGORITHM - Find Object Initial position of camera – Looking straight. Check: can camera see the object ? Yes- Object found so align robot to object Cannot see the object Look around for object Look Around for object Rotate the camera using the servo motor mounted below the camera starting at 0 degrees up to 180 degrees. Scan for the colored object at small intervals of time during the motion. Check: Did we detect the object during rotation? Yes - Object found so align robot to object No – Move Minimum distance and start over


ALGORITHM - FIND Nest : ALGORITHM - FIND Nest Initial Position – Object found and is gripped. Check: can camera see the required nest ? Yes- Nest found so align robot to move to nest Cannot see the nest Look around for the right nest. NEST FOUND: Move Robot till the confidence reading for the nest color is very high and proximity sensor can detect the wall, at this stage we are ready to drop the object i.e. the nest is reached.


Outline: Outline Problem Analysis Hardware Basics Software Architecture Basics Find Object or Find Nest Get Object Deposit Go to Nest Avoid Obstacle Conclusion


Proximity Sensor: Proximity Sensor Infrared Sensor Used to detect an object in the following way : Sends an infrared signal If an object is present this signal is reflected of an object. A receiver then collects this signal and detects the presence of the object. Range from 10 centimeters (4") to 180 centimeters (72”)


Touch Sensor: Touch Sensor Function of a touch sensor : Detects contact with objects in the robot's environment. If its button is pressed or not pressed


Light Sensor: Light Sensor A light sensor measures the amount of light that it sees. It reports the amount of light as a number between 0 (total darkness) and 100 (very bright). When the light sensor is over the white paper, it reads a value of 50. When it is over the black paper, it reads a value of 33.


State Diagram - Get Object: State Diagram - Get Object


ALGORITHM - Get Object: ALGORITHM - Get Object If (proximity sensor sees object as 10 cms away) {    Travel 10 cms    Turn on Gripper Motor    Continue Step 2 until touch sensor activated    Turn off Gripper Motor } Return (Program control to Planner Stage )


ALGORITHM - Deposit Object: ALGORITHM - Deposit Object If (in nest) {    Turn on Gripper Motor until touch sensor activated.    Turn on reverse engine for one second    Rotate 180 degrees } Return (program control to planner stage)


Outline: Outline Problem Analysis Hardware Basics Software Architecture Basics Find Object or Find Nest Get Object Deposit Go to Nest Avoid Obstacle Conclusion


Rear wheel drive : Rear wheel drive Has 4 wheels 2 motors attached to the back wheels. 2 rotation sensors attached to the back wheels. The wheels are attached with axels which also goes through the rotation sensor.


DC Motor: DC Motor Converts electrical to mechanical energy


Rotation Sensor: Rotation Sensor The rotation sensor is a vital component in measuring distance traveled and angles turned of the robot To measure the distance traveled by the robot, the sensor counts the internal clicks turned by the wheel axle It takes 16 clicks for the sensor to turn 360 degrees, which is one complete turn of the wheel axle.


ALGORITHM- GOTO NEST: ALGORITHM- GOTO NEST Move to the nest Use motors to turn both wheels at the same speed to go forward use the rotation sensor to measure the distance traveled If there is an obstacle in path of the nest, avoid the obstacle Turn right by turning left motor forward and right one in reverse use rotation sensor to count 4 clicks – 90 degrees move forward slightly, turn left to come back on track to the designated nest


Current Status: Current Status We have built the basic robot body Tested the camera and other sensors being used and calibrated them. Completed the design Our next task is the implementation phase .


FUTURE ENHANCEMENTS: FUTURE ENHANCEMENTS Our robots design can be enhanced to allow mapping and navigation of uncharted terrain. Use IR signals to allow two robots to communicate Develop teams of robots that can cooperate.


REFERENCES: REFERENCES Robotic Explorations - By Fred G. Martin Building Robots with Lego Mindstorms - By Ferrari & Hempel Lego Mindstorms By David Astolfo Gripper building logic – www.ozbricks.com Other valuable resources www.handyboard.com www.cs.cmu.edu