McCrawler An Autonomous OmniDirectional SixLegged Robot

1
McCrawler An Autonomous Omni-Directional
Six-Legged Robot

• By
• Heike Sichtig Jie Yuan

2
Introduction

• As Rod Brooks, one of the early critics of
  mainstream AI stated
• Intelligent behavior can be generated without
  explicit abstract reasoning of the kind that
  symbolic AI proposes. Intelligence is an emergent
  property of certain complex systems.
• McCrawler will be a Real world application
  with a reactive architecture, exhibiting
  intelligent behavior in the eye of the
  beholder. Obstacle Avoidance can be achieved
  rather simply, using such architecture.

3
McCrawler

• Able to navigate on almost any surface
• Mars explorations
• Obstacles Avoidance using advanced
  telecommunication technologies.
• Infrared technology
• Some parts of McCrawler are similar to living
  creatures, like spiders

4
McCrawler Design

• Body
• a strong, outer layer that protects the robot's
  electronics parts and batteries
• Brain
• the robots brain is embedded in the body,
  containing an integrated circuit (C324C) to
  process information
• Eyes and other "senses
• IR sensors and instruments that give the robot
  information about its environment        
• Six legs
• a way to extend its reach and react to the
  environment, also part for mobility
• Energy
• batteries
• Future communications
• mobile communication like Blue tooth or WIFI  

5
Reactive Architectures

• Simple architecture that can produce complex
  behavior
• No representations of the environment or complex
  problem solving
• Can use dedicated, parallel hardware
• Fast (real-time) response to changes in the
  environment

6
Reactive Architectures

• Disadvantages
• Fixed response to a given situation
• All responses must be defined in advance
• Not able to cope with novel situations for which
  they dont have a predefined behavior
• Not able to solve some problems at all

7
Why A six-legged robot?

• Two main types of mobile robots, wheeled robots
  and legged robots.
• In this project we choose to use the legged
  robots. The fundamental issue for legged robots
  is walking. 7
• If the robot is to move around in an unknown
  environment, it should at least be able to handle
  coordination of joint movements in spite of
  relatively small perturbations, obstacle
  avoidance and stability.

8
How to install

• Mechanical Assembly
• Gear Mechanism
• Slider Crank Mechanism

9
How to install

• Soldering the PC Board
• Emission Section
• Projects the light for detecting obstacles, LED
  emits red light in the form of pulses)
• Sensing Section
• Light emitted in form of pulses reflects on a
  phototransistor when it hits the obstacle
• Voltage Amplification Section
• A large voltage is needed to control the motors
• Rectification Circuit Section
• Signals in a pulse form, are rectified through
  this section
• One Shot Timer Section
• This section makes a timer ,so the robot will not
  return immediately
• Signal Reversal Section
• Reverse Signal to make the motor run reverse

10
How to install
Electronic Circuit -Block Diagram-
11
Research Question

• Will McCrawler, a simple reactive agent, be able
  to successfully avoid most collisions in a
  dynamic environment, to be deployed for
  exploratory missions, like robots on planet Mars?
• Agent architecture
• Sensing module
• Execution module
• No internal states
• Environment
• Dynamic, i.e. the world can change independently
  from the agent's action
• Agents goals
• static and simple Avoid obstacles

12
Evaluation Methodology

• Circuit Board Evaluation
• Real Environment
• Functionality Study
• Behavior Study
• Performance Study

13
Evaluation Methodology- Functionality Study

• After assembly, test the robots mobility in a
  real environment.
•  ü Structured Environment
• Record whether the parts of the robot act
  according to the design of the circuit board
• Check mobility of robot

14
Evaluation Methodology- Behavior Study

• Test whether the robot behaves according to its
  description
•  ü   Obstacle Avoidance
• Wall
• Corner
• High/ Low Speed Moving object
• Throw a Ball
• Avoid a person
• Avoid a dark back pack
•  ü   Power Consumption
• How long will McCrawler last

15
Evaluation Methodology- Performance Study

• Define various goals to test desired performance
  of McCrawler
• ü Goal 1
• Make turn when reaching wall
• ü Goal 2
• Make move when encountering a corner 
• ü Goal 3
• Avoid moving object, such as ball
• Make turn when reaching wall

16
Results

• Circuit Board Evaluation
• Circuit test was positive
• Real Environment
• Functionality Study
• Slider Crank Mechanism acted according to manual
• Robot moved when circuit was connected
• Behavior Study
• Robot successfully avoided most walls and corners
• Reflectivity of objects seems to be a major point
• Robot sometimes avoided moving ball
• Robot failed avoiding back pack
• Robot most of the time avoided human being
• Performance Study
• All goals were successfully accomplished

17
Discussion

• McCrawler is functional, but has many limitations
• Reflectivity is a big factor
• When objects to dark, McCrawler is not able to
  respond
• Improvements in the sensing module are needed
• Agent Architecture
• Reactive behavior is very limited
• Needs capability of storing internal state for
  deployment in real world explorations
• McCrawler as such is not the best choice as an
  exploration agent, when encountering a problem
  (dark rock like a bag pack) .
• Future research will enhance McCrawler

18
Future Research - Sensors

• Beware of stairs
• Be able to make smart turns depending on the
  situation, such as learns to associate signals
  coming from the reflectance of light sensors with
  the corresponding motor actions.

19
Future Research - wireless communication method

• Bluetooth (like sony remote control car?)
• short Range
• Low power consumption
• 802.11b
• IR
• Hardware control or software
• Ad hoc network or Access point?

20
Future Research - To enable multiple mobile
agent communication

• Proactive routing protocols
• high overhead , poor scalability, but good
  reliability, low latency
• Demand-based, reactive route discovery
• low routing overhead, increased latency
• Preemptive route discovery
• support mobility still remains a point of
  research, especially in the context of
  time-constrained applications
• MARP for Mobile Wireless Ad Hoc Networks
• MARP performs well when the velocity of nodes is
  below 45 m/s.

21
Future Research- more Travel Algorithm

• Be able to navigate purposefully from a start
  location to a target, needs two basic
  requirements sensing and reasoning.
• Repeatedly travel back and forth between two
  points, finding a shortest path in the
  environment
• Randomized robot exploration

22
Future Research - Adaptive Light Compass

• Robot navigation by light
• Like ants homing, use a biologically inspired
  orientation mechanism, an adaptive light compass
  used for homing
• Needs to be equipped with infrared and ambient
  light sensors.

23
Future Research - Human pet robot interaction

• Like study conducted by MERL found out that human
  subjects with a robot designed to mimic human
  conversational gaze behavior in conversation.
• We can conduct study about human subjects
  reaction with robot designed to mimic Pet
  behavior such as make noises, express feelings
  such as surprise or fatigue, following, seeking
  attention, roll-over .