Robot Anatomy

W.S.Harwin

January 2016

Index
Next Lecture
Previous Lecture
Additional material

Robot Anatomy

Extras at http://www.reading.ac.uk/~shshawin/LN

Key components

Cerebral cortex

Sensors

  • Passive sensors simply listen in on what is happening,
  • Active sensors transmit into the environment and measure response.

Animals have both interesting sensing capabilities and the ability to process (and exploit) the data

  • Sensors can produce large amounts of data for interpretation

Heat: Sidewinder or pit viper

Crotalus cerastes.

cont

  • Infra red cameras (need vision recognition software)
  • PIR detectors (sensitive but indiscriminate)

Smell: Noctuid moth and Cape Buffalo

Noctid moth

Syncerus caffer.

Sense of smell

  • Artificial nose
  • Polution sensors (not much better than a canary)

Low frequency EM: - Electrical organs

Psiticus
Platypus

Low frequency electromagnetic sensors

  • Electrostatic fields
    • proximity sensors
    • Proximity switch on robots, lights etc
    • Output is usually binary
  • Radar
    • Good for velocity detection
    • Air traffic control etc humans still main source of intelligence

Pressure: Whale shark (Rhincondon)

Inertial navigation: Fly halters

Inertial navigation: (what we have)

  • Rate gyroscopes for angular velocity
  • Accelerometers for linear accelerations
  • Good candidates for control systems, but drift is a problem
  • Used in consumer products such as Wiimote and Segway

Touch: Crayfish

palinuridae panulirus ornatus (Rock Lobster (Crayfish))

Touch: Star-nose mole (Condylura cristata)

Condylura cristata (Star nose mole)

Touch

  • Robot bump sensors
  • Robot whiskers
  • Both are passive devices, both give binary information

Taste: Nomad

Visible light: White bellied sea eagle

Visible light: Cats

Domestic cat Felis catus
Panthera Leo

Visible light

  • Video Cameras
  • Light intensifiers
  • Much interest in vision processing, can track objects in sequence of pictures
  • Simplistic algorithms eg edge detection

Canny edge detection

Lena Soderberg

Low freq EM Magnetic: swallow

Quasi-static EM fields.

Hirundu rustica

magnetic field

  • Earths field
    • Navigation
    • Problems indoors where it gets distorted
  • Self generated magnetic field
    • Good for position tracking eg for VR applications

Infrasound

Loxodonta africana

Ultrasound

  • Bats use continuous modulation (CM), frequency modulation (FM) or a mixture of both (hockystick)
  • The UK Pipistrelle uses CM at frequencies of 39kHz 45kHz 55kHz depending on species
  • Due to attenuation the range is probably around 20 metres for bats
  • Myotis lucifugus uses scans from 70KHz-33KHz in .2 seconds
  • echo location can be used in humans to aid navigation (e.g. Daniel Kish)

Myotis lucifugus (FM 70KHz-33KHz in .2 seconds) Brown bat

Big brown bat (Eptesicus fuscus) flying in an obstacle test with rows of vertically hanging plastic chains. The bat's head, ears, and sound beam are aimed toward the open path to the front. Credit: James Simmons

Tursiops truncatus (bottle nose dolphin)

Sound

  • Ping and listen devices discard 90% of information
  • Military interest, but again humans used as the intelligence

Rotary angle sensor

Passive and active sensors

Transducer/sensor quantity
strain gauges Force. Change in resistance implies distortion of material implies force
Linear variable differential transformerDistance. Via magnetic coupling to a coil
Hall effect Magnetic field. Electrons in semi-conductor move under the magnetic field to a gate
Laser range finder distance. Time of flight of photons from transmitter, reflected off object
ultrasound modules distance. Time of flight of sound.

GPS receivers position. Time of flight of timed RF photons from satelite to the receivers
Accelerometers linear accelerations. Measure acceleration via force (Newton $F=ma$) see strain-gauges
Gyrometer angular accelerations. Several mechanisms including 'tuning forks' and optical fibres. Old style used spinning gyroscopes
Tachometers angular velocity

Kinematics part 1

  • Degrees of freedom to locate (3DoF) and orientate (3DoF) a body
  • Redundant Dof allows reach around (human arm from shoulder to wrist has approx 6DoF, but then we can move the trunk to reach around)
  • Revolute and prismatic joints
  • Active (contribute to DoF) and Passive
  • Serial chains, parallel chains and combination linkages tend to trade off considerations such as reach, range of movement, weight and strength.

Lower joint pairs

Name (Symbol) DoFcontains type
Revolute (R) 1R planara hinge
Prismatic (P) 1P planaran aerial
Helical (H) 1R+P 3Da screw
Cylindrical (C) 2R+P 3D
Spherical (S) 33R 3Dshoulder
Sliding/Flat (F) 3R+2P 3D

Revolute (R) and Prismatic (P)

Serial chains (Cartesian)

Gantrys PP

Serial chains (revolute)

JCB RRRR

SCARA- PRR or RRP

Cylindrical robot RPP

Parallel chains (Stewart platform)

Passive and active joints

  • Only active joints contribute to degrees of freedom
  • Passive degrees of freedom needed to remove motion constraints.
  • True Stewart platform is (PPPPPP)
  • Grueblers equation and Kutzbach criterion

Delta robot Serial/Parallel (RR)(RR)(RR)

Questions?