Mobile & Service Robotics Mobile & Service Robotics Sensors - - PowerPoint PPT Presentation

Mobile & Service Robotics Mobile & Service Robotics

Sensors for Sensors for Robotics Robotics – 1 Sensors for Sensors for Robotics Robotics 1

An Example of robots with their sensors

2 Basilio Bona – Robotica 03CFIOR ‐ 2011

Another example

Omnivision Camera (360°) Pan-Tilt-Zoom (PTZ) camera Sonars IMU= Inertial Measurement Unit Laser Scanner Encoders inside differential wheels Bumpers

3

Passive support wheel

Basilio Bona – Robotica 03CFIOR ‐ 2011

Definition

A sensor is a device that produces a measurable response to a change in a physical condition (such as temperature) or to a change in a chemical concentration Usually commonly used sensors convert the physical quantity Usually commonly used sensors convert the physical quantity into a signal which can be measured electrically The sensors are classified according to the following criteria: The sensors are classified according to the following criteria:

1. Primary Input quantity (aka measurand) 2 Transduction principles 2. Transduction principles 3. Measured property (as temperature, flow, displacement, proximity, acceleration, etc.) proximity, acceleration, etc.) 4. Material and technology 5 Application 5. Application

4 Basilio Bona – Robotica 03CFIOR ‐ 2011

Sensors types

Proprioceptive sensors (PC)

They measure quantities coming from the robot itself, e.g., motor speed, wheel loads, robot heading, battery charge status, etc.

Exteroceptive sensors (EC)

They measure quantities coming from the environment: e.g., walls distance, earth magnetic fields, intensity of the ambient light, obstacle positions, etc.

Passive sensors (SP)

They use the energy coming from the environment

Active sensors (SA)

They use the energy they produce and measure the reaction of the environment (better performance, but may influence the environment)

5 Basilio Bona – Robotica 03CFIOR ‐ 2011

A l S th ti i bl d id th

Sensors types

Analog Sensors: they measure continuous variables and provide the information as a physical reading (mercury thermometers and old style voltmeters are good examples of analog sensors) style voltmeters are good examples of analog sensors) Digital Sensors: they measure continuous or discrete variables, but the provided information is always digital, i.e., discretized p y g , , Continuous Sensors: although the name is somehow misleading, Continuous Sensors: although the name is somehow misleading, continuous sensors (analog or digital) provide a reading that is on a continuous range, as opposite to ON/OFF sensors Binary Sensors : they give only two levels of information ON/OFF or YES/NO: a lamp that switches on when a certain temperature level is attained, is an analog binary sensor

6 Basilio Bona – Robotica 03CFIOR ‐ 2011

Sensors classification

Category Sensors Type Tactile senso s/ p o imit Contact sensors (on/ off), bumpers EC - SP Proximity sensors EC SA Tactile sensors/ proximity sensors y (inductive/ capacitive) EC - SA Distance sensors (inductive/ capacitive) EC - SA Active wheel sensors Potentiometric encoders PC - SP Optical, magnetic, Hall-effect, inductive, capacitive encoders, PC - SA inductive, capacitive encoders, syncro and resolvers PC SA Heading sensors with respect to Compasses EC - SP Gyroscopes PC SP g p a fixed RF Gyroscopes PC - SP Inclinometers EC – SP/ A GPS (outdoor only) EC – SA Absolute cartesian sensors Optical or RF beacons EC – SA Ultrasonic beacons EC – SA Refelctive beacons EC – SA Refelctive beacons EC SA

7 Basilio Bona – Robotica 03CFIOR ‐ 2011

Sensors classification

Category Sensors Type Reflective sensors EC - SA Active distance sensors (active ranging) Ultrasonic sensors EC - SA Laser range finders, Laser scanners EC - SA (active ranging) Optical triangulation (1D) EC - SA Structured light (2D) EC - SA Motion and velocity sensors (speed relative to fixed or mobile objects) Doppler radar EC - SA Doppler sound EC - SA Vision sensors: distance from stereo vision, feature analysis, segmentation object CCD and CMOS cameras EC - SA Integrated packages for visual ranging EC - SA segmentation, object recognition g g Integrated packages for object tracking EC - SA

8 Basilio Bona – Robotica 03CFIOR ‐ 2011

Sensor characteristics

D i Dynamic range Resolution Linearity

Bandwidth or frequency Transfer function

Reproducibility/precision Accuracy Systematic errors Systematic errors Hysteresis Temperature coefficient Temperature coefficient Noise and disturbances: signal/noise ratio C t Cost

9 Basilio Bona – Robotica 03CFIOR ‐ 2011

Sensor characteristics

Dynamic range

• Ratio between lower and upper measurement limits, expressed in

dB dB

• Example: voltage sensor min=1 mV, max 20V: dynamic range 86dB
• Range = upper limit of dynamic range
• Range = upper limit of dynamic range

Resolution

• Minimum measurable difference between two values
• Resolution = lower limit of dynamic range
• Digital sensors: it depends on the bit number of the A/D converter
• Example 8 bit=25510 range 20 V ‐> 20/255 = 0.08

Bandwidth

• Diff

b t d l f i

• Difference between upper and lower frequencies
• Large bandwidth means large transfer rate
• Lower bandwidth is possible in acceleration sensors

Lower bandwidth is possible in acceleration sensors

10 Basilio Bona – Robotica 03CFIOR ‐ 2011

Accuracy and precision

11 Basilio Bona – Robotica 03CFIOR ‐ 2011

Accuracy and Precision

P i i R bili R d ibili Precision = Repeatability = Reproducibility

Precise but Accurate but t i not accurate not precise Precise and accurate Not accurate and not precise

12 Basilio Bona – Robotica 03CFIOR ‐ 2011

Noise Noise

13 Basilio Bona – Robotica 03CFIOR ‐ 2011

Noise

All sensors are subject to noise since due to random All sensors are subject to noise, since, due to random fluctuations or electromagnetic interference, they add to the measured signal an undesired component that cannot be measured signal an undesired component that cannot be precisely known If th i i ll th th t fl t ti d If the noise is smaller than the measurement fluctuations and the noise introduced by the electronic components, it is not influent influent On the contrary it can degrade the entire chain plant‐sensor‐ controller and make it unusable

14 Basilio Bona – Robotica 03CFIOR ‐ 2011

Noise

Noise is often spread on a large frequency spectrum and many noise sources produce the so‐called white noise, where the power spectral density is equal at every frequency The noise is often characterized by the spectral density of the noise Root Mean Square (RMS), given as Since it is a density, to obtain the RMS value one shall integrate

/ V Hz

y, g the spectrum density in the frequency band of interest. This type of distribution adds to the measure an error term that is proportional to the square root of the bandwidth of the measuring system

15 Basilio Bona – Robotica 03CFIOR ‐ 2011

Noise types

Noise are of many types; these include Noise are of many types; these include Shot noise Thermal noise Thermal noise Flicker noise Burst noise Avalanche noise To know the noise type is important for modeling purposes

16 Basilio Bona – Robotica 03CFIOR ‐ 2011

Shot noise

Shot noise, often called quantum noise, is always associated to random fluctuations of the electric current in electrical conductors, due to the current being carried by discrete charges (electrons) whose number per unit time fluctuates d l randomly This is often an issue in p‐n junctions. In metal wires this is h l l b d d l much less important, since correlation between individual electrons remove these random fluctuations Shot noise is distinct from current fluctuations in thermal equilibrium, which happen without any applied voltage and ith t t fl i Th th l without any average current flowing. These thermal equilibrium current fluctuations are known as thermal noise Th h t i t i fl t The shot noise spectrum is flat

17 Basilio Bona – Robotica 03CFIOR ‐ 2011

Thermal noise

Thermal noise, also called Johnson–Nyquist noise, is the electronic noise generated by the thermal agitation of the h i ( ll h l ) i id l i l charge carriers (usually the electrons) inside an electrical conductor at equilibrium, which happens regardless of any applied voltage applied voltage Thermal noise is approximately white pp y With good approximation the amplitude of the signal has a b b l d f Gaussian probability density function

18 Basilio Bona – Robotica 03CFIOR ‐ 2011

Flicker noise

Flicker noise, also called 1/f noise or pink noise is characterized by a frequency spectrum such that the power spectral density is inversely proportional to the frequency It is always present in active components of electronic circuits and in many passive ones It is proportional to the current amplitude, so if the current is sufficiently low, the thermal noise will predominate

19 Basilio Bona – Robotica 03CFIOR ‐ 2011

White noise

White noise is a random signal (or process) with a flat power spectral density The signal contains equal power within a fixed bandwidth at any center frequency An infinite‐bandwidth white noise signal is a purely theoretical construction The bandwidth of white noise is limited in practice by the mechanism of noise generation, by the transmission medium and by finite observation capabilities A random signal is considered “white noise” if it is observed to have a flat spectrum over the widest possible bandwidth White noise is often used for modeling purposes g p p

20 Basilio Bona – Robotica 03CFIOR ‐ 2011