EC7010:

Sensors for robots

Dr. Lilantha Samaranayake
Department of Electrical & Electronic Engineering
Faculty of Engineering
University of Peradeniya
[email protected], +94812393426
Robot Sensors

 Why do Robots need sensors?

 What can be sensed?
 What sensors are out there?
 What can they do?
 How much do they cost?
 How easy are they to use?
 Characterizing sensor performance
Why do Robots need sensors?

 Provides “awareness” of surroundings

 What’s ahead, around, “out there”?
 Allows interaction with environment
 Robot lawn mover can “see” cut grass
 Protection & Self-Preservation
 Safety, Damage Prevention, Stairwell sensor
 Gives the robot capability to goal-seek
 Find colorful objects, seek goals
 Makes robots “interesting”
Sensors - What can be sensed?
 Light
 Presence, color, intensity, content (mod), direction
 Sound
 Presence, frequency, intensity, content (mod), direction
 Heat
 Temperature, wavelength, magnitude, direction
 Chemicals
 Presence, concentration, identity, etc.
 Object Proximity
 Presence/absence, distance, bearing, color, etc.
 Physical orientation/attitude/position
 Magnitude, pitch, roll, yaw, coordinates, etc.
Sensors - What can be sensed?

 Magnetic & Electric Fields

 Presence, magnitude, orientation, content (mod)
 Resistance (electrical, indirectly via V/I)
 Presence, magnitude, etc.
 Capacitance (via excitation/oscillation)
 Presence, magnitude, etc.
 Inductance (via excitation/oscillation)
 Presence, magnitude, etc.
 Other Things?
What sensors are out there?

 Feelers (Whiskers, Bumpers) – Mechanical

 Photoelectric (Visible) – Active & Passive
 Infrared (light) – Active & Passive
 Ultrasonic (sound) – Active & Passive
 Sonic – Active & Passive
 Resistive/Capacitive/Inductive – Active & Passive
What sensors are out there?

 Visual – Cameras & Arrays (Active & Passive)

 Color Sensors (Active & Passive)
 Magnetic (Active & Passive)
 Orientation (Pitch & Roll)
 GPS (location, altitude)
 Compass (orientation, bearing)
 Voltage – Electric Field Sensors
 Current – Magnetic Field Sensors
 Chemical – Smoke Detectors, Gas Sensors
A Closer Look
Sensors – Feelers
 Whiskers
 Piano wire suspended through conductive “hoop”
 Deflection causes contact with “hoop”
 Springy wire that touches studs when deflected
 Reaches beyond robot a few inches
 Simple, cheap, binary output

 Bumpers & Guards

 Impact/Collision sensor, senses pressure/contact
 Microswitches & wires or framework that moves
 Simple, cheap, binary output, easy to read
Feelers - Whiskers
Feelers - Bumpers & Guards

From Kevin Ross’s “Getting Started Article (SRS Website)

Sensors – IR
 Active (emitting)
 Oscillator generates IR reflections off objects
 Filtered receiver looks for “reflections”
 Pulses may be encoded for better discrimination
 Typically frequencies around 40KHz
 Doesn’t work well with dark, flat colored objects

 Passive (sensor only)

 Pyro-electric (heat sensor)
 Look for IR emissions from people & animals
 Used in security systems & motion detectors
Infrared - Active
Active IR sensor specifications
 Sensor type = Reflective IR
 IR detector = Panasonic PNA4602M
 IR LED type = Narrow focus 10º
 I/O required = 3 digital lines: 2 outputs, 1 input
 Range = Approximately 4“ to 26"
 Input voltage = 5vdc regulated @ 8mA
 PC board size = 2.3" x .75"
Linear Array IR Range Sensors
 Sharp GP2Dxx (one of many)
 ~4 to 30cm Range
 Fixed Range with Discrete Output
 Analog or Digital Output
 Easy to Use
Laser Range Sensors
 USB Interface
 240º Field of View
 0.36º Angular Resolution
 10Hz Refresh Rate
 20mm to 4m
 Price can be from $100 to $3000
Passive IR – Pyro-Electric

The Model 442-3 IR-EYE is a Lithium Tantalate pyroelectric parallel opposed

dual-element high-gain detector with complete integral analog signal
processing.
Sensors – Ultrasonic
 Active
 Emit pulses & listen for echos
 Times round trip sound travel (~1ft/mS)
 Reaches far fairly beyond robot (inches to 30-50’)
 Relatively simple, not cheap, analog output
 Directional; not everything reflects sound well

 Passive (listens only)

 Sensor listens for ultrasonic sounds
 Electronics may translate frequency or modulation
 Software may perform signal analysis (FFTs, etc.)
Ultrasonic - Active

Price increases

Visit http://www.acroname.com for more information about these & other products.
 Sensors – Ultrasonic (cont.)
 Passive - Beacons & Sensors
 Beacons listen: RF command to broadcast
 Send light & sound pulses
 Robot looks & listens for each beacon
 Light pulse starts timer, sound pulse stops it
 Robot knows location of each beacon
 Compass on robot provides its orientation
 Robot computes distance, measures bearing
 Robot can then compute its location

(Speed of Light=3x108 m/s, Speed of Sound=340 m/s)

Ultrasonic - Passive

Beacon3
Beacon1
Sound Pulse
Sound Pulse
Light Pulse

Light Pulse

Robot
Sound Pulse
BEACON ROBOT Light Pulse
RF Receiver RF Transmitter
Beacon2
IR Emitter IR Receiver
Ultrasonic Emitter Ultrasonic Receiver
Sensors – Sonic (Acoustic)
 Active
 Emit pulses & listen for echos
 Times round trip sound travel (340 m/s)
 Reaches far fairly beyond robot (10 – 15 m)
 Relatively simple, not cheap, analog output
 Directional, not everything reflects sound
 Noisy!!!!

 Passive (sensor only)

 Sensor listens to ambient sounds
 Filters or scans selected frequencies
 ADC measures conditioned signal amplitude
 CPU performs signal analysis on what it hears
Sonic (Acoustic) - Passive

TI (Burr-Brown) UAF42 Universal Active Filter

http://focus.ti.com/lit/an/sbfa002/sbfa002.pdf
Sensors – Resistance

 Passive (sensor only)

 Measures elec. resistance between objects
 Measure sensor that varies resistance
 Use absolute or differential readings
 Other ideas?
Sensors – Capacitive
 Passive
 Really doesn’t work (Needs excitation)

 Active (emitting)
 Generate AC or DC voltage
 Apply to external environment
 Measure current to determine Resistance
 Short range applications
Sensors - Capacitive
 Sensors – Inductive
 Passive
 Really doesn’t work (Needs excitation)

 Active (emitting)
 Current flows through inductor
 Magnetic field mostly ignores non-metals
 Inductance changes with metallic proximity
 Short range applications
 Sensors - Inductive
 Passive
 Really doesn’t work (Needs excitation)

 Active (emitting)
 Metals affect sensor
 Current flows through inductor
 Magnetic field mostly ignores non-metals
 Inductance changes with metallic
proximity
 Short range applications (~cm or mm)
Sensors – Visual
 Active (emitting)
 Camera with field of view illumination
 Looks for particular reflections
 Filter removes non-significant light sources

 Linear array senses single axis of motion

 Passive (camera only)

 Scans field of interest
 Looks for objects, artifacts, features of interest
 Processes digital data to simplified interpretation
Sensors – Visual

 CMUCam
 Linear Optical Array
Sensors – Color
 Active (emitting)
 Selective field illumination (specific color(s))
 Sensor filter removes extraneous light sources
 Output can be analog (prop.) or digital (on/off)

 Passive (sensors only)

 Different sensors for different colors
 Color filter removes extraneous light sources
 Output can be analog (prop.) or digital (on/off)
Sensors – Color
http://robotroom.com/ColorSensor.html
Sensors – Orientation

 Rate Gyroscope
 Output proportional to angular speed
 Integrate to get position
 Differentiate to get acceleration

 DC Accelerometer
 Output proportional to sine of vertical angle
Sensors – Motion
Rate Gyro – Silicon Sensing Systems
Servo Accel – Sensorland.com
Sensors – Position/Location

 Wheel Encoders
 Relative position & motion
 Integrate/Differentiate for other parameters

 Global Positioning System

 Absolute position/location on earth
 Local differential error correction
 Integrate/Differentiate for other parameters
Wheel Encoders
Sensors – Compass (Orientation)

 Track bearing angle & distance to determine

position
 Sensors – Voltage
 Passive – Senses electric field
 Fluke Electric Field Sensors
 Sensors – Current
 Series measurement with Hall Effect device
 Current loop (coil), then amplified
 Magnetoresistive (Wheatstone bridge)
Sensors – Current
Sensors – Current
 Sensors – Chemical
 Passive (sensors only)

 Active (optical emitter/photo sensor)

 Sensors – Chemical
 Smoke Detectors - Cheap, readily available, $5
 Oxygen concentration sensors - CO, H4S, CH4, pricey
 Characterizing sensor performance
 Dynamic Range
 Ratio between upper and lower limits (usually in decibel)
 Power measurements (1 mW to 20 W)
10log(20/0.001) = 43dB
 Voltage measurements (1 mV to 20 v)
20log(20/0.001) = 86dB
 I.e. F log Max Mix , where F=10 for power entities and F=20 for no-power
entities
 Characterizing sensor performance contd.,
 Range
 Upper limit of measurements
 Resolution
 minimum difference between two values
 often lower limit = resolution
 Linearity
 Variation of output as a function of input
 Ideally Y = aX implies Y = a(X1 + X2)
 Bandwidth/Frequency
 Speed of response, delay
 Characterizing sensor performance contd.,
 Sensitivity
 minimum input change to result in output change
 Cross sensitivity
 Variation with other changes such as temperature
 Error/accuracy
 Difference between actual value and generated value
 accuracy = 1−(|m−v|/v) where m is the measured value and v is the true
value
 Characterizing sensor performance contd.,
 Most sensors generates measurements that are
contaminated by noise.
 Systematicnoise: errors that could be modelled for
example through calibration
 Random noise: errors that cannot be predicted.
Typically modelled in a probabilistic fashion
 Precision: reproducibility of measurements
precision = range / 
Where  is the standard deviation
Sensors – Conclusion

 Sensors
 provide a way of simulating “aliveness”
 give robots environmental awareness
 provide of means of human protection
 help robot preserve itself
 enable goal seeking
 enable closed-loop interaction
 make robots interesting
 can make programming “challenging”