CY 02CFIC CFIDV RO OBOTIC CA 01 MOBILE ROBOTICS Sensors An Introduction Basilio Bona DAUIN Politecnico di Torino Basilio Bona DAUIN Politecnico di Torino 001/1
CY CA 01CFIDV 02CFIC OBOTIC RO An Example Basilio Bona DAUIN Politecnico di Torino 001/2
An Example Omnivision Camera (360 ) CFIDV 02CFIC CY RO OBOTIC CA 01 Pan-Tilt-Zoom (PTZ) camera IMU=Inertial Measurement Unit Sonars Laser Scanner Encoders Bumpers Passive wheel Basilio Bona DAUIN Politecnico di Torino 001/3
Sensors Type CFIDV 02CFIC CY RO OBOTIC CA 01 Proprioceptive sensors (PC) They measure quantities coming from the system, e.g., motor speed, wheel load, heading of the robot, battery charge status, etc. Exteroceptive sensors (EC) They measure quantities coming from the environment al robot: e.g., wall distance, magnetic fields, intensity of the ambient light, obstacle position, etc. Passive sensors (SP) They use the energy coming from the environment (not to be confused with the energy required to move) Active sensors (SA) Emit their proper energy and measure the reaction of the environment Better performance, but may influence the environment Basilio Bona DAUIN Politecnico di Torino 001/4
Sensors Type 02CFIC CY CFIDV RO OBOTIC CA 01 Analog Sensors Digital Sensors Continuous Sensors Binary Sensors (ON/OFF) Basilio Bona DAUIN Politecnico di Torino 001/5
Sensors Classification CFIDV 02CFIC CY RO OBOTIC CA 01 Category Sensors Type Tactile sensors/proximity s/p o sensors Active wheel sensors Heading sensors with respect to a fixed RF Absolute cartesian sensors Contact sensors (on/off), bumpers Proximity sensors (inductive/capacitive) Distance sensors (inductive/capacitive) Potentiometric encoders Optical, magnetic, Hall-effect, inductive, capacitive encoders, syncro and resolvers Compasses EC - SP PC - SP PC - SA EC - SP Gyroscopes PC - SP Inclinometers GPS (outdoor only) Optical or RF beacons Ultrasonic beacons EC SP/A EC SA EC SA EC SA Refelctive beacons EC SA Basilio Bona DAUIN Politecnico di Torino 001/6
Sensors Classification CFIDV 02CFIC CY RO OBOTIC CA 01 Category Sensors Type Active distance sensors (active ranging) Motion and velocity sensors (speed relative to fixed or mobile objects) Vision sensors: distance from stereo vision, feature analysis, segmentation, object recognition Reflective sensors Ultrasonic sensors Laser range finders, Laser scanners Optical triangulation (1D) Structured light (2D) Doppler radar Doppler sound CCD and CMOS cameras Integrated packages for visual ranging g Integrated packages for object tracking Basilio Bona DAUIN Politecnico di Torino 001/7
Sensor Characteristics CFIDV 02CFIC CY RO OBOTIC CA 01 Transducer = Sensor Dynamic range and range Resolution Linearity Dynamic range Bandwidth or frequency Transfer function Reproducibility/precision Accuracy Systematic errors Hysteresis Temperature coefficient Noise and disturbances: signal/noise ratio Basilio Bona DAUIN Politecnico di Torino 001/8
Sensor Characteristics CFIDV 02CFIC CY RO OBOTIC CA 01 Dynamic range Ratio between lower and upper limits expressed in db Example. Voltage sensor min=1 mv, max 20V: dynamic range 86dB Range = upper limits Resolution Minimum measurable difference between two values Lower limits of dynamic range = resolution Digital sensors: it depends on the bit number of the A/D converter Example 8 bit=255 10 range 20 V -> 20/255 Bandwidth Large bandwidth means large transfer rate Lower bandwidth is possible in acceleration sensors Basilio Bona DAUIN Politecnico di Torino 001/9
CY CA 01CFIDV 02CFIC OBOTIC RO Accuracy and Precision Basilio Bona DAUIN Politecnico di Torino 001/10
Accuracy and Precision Precision = Repeatability = Reproducibility 02CFIC CY RO OBOTIC CA 01CFIDV Precise but not accurate Not accurate and not precise Accurate but not precise Precise and accurate
ROBOTICA 01CFIDV 02CFICY Noise Basilio Bona DAUIN Politecnico di Torino 001/12
Noise Types 02CFIC CY CFIDV CA 01 OBOTIC RO Shot noise Thermal noise Flicker noise Burst noise Avalanche noise Basilio Bona DAUIN Politecnico di Torino 001/13
CY CA 01CFIDV 02CFIC OBOTIC RO Shot noise Basilio Bona DAUIN Politecnico di Torino 001/14
CY CA 01CFIDV 02CFIC OBOTIC RO Thermal noise Basilio Bona DAUIN Politecnico di Torino 001/15
CY CA 01CFIDV 02CFIC OBOTIC RO Flicker Noise Basilio Bona DAUIN Politecnico di Torino 001/16
CY CA 01CFIDV 02CFIC OBOTIC RO Flicker Noise Basilio Bona DAUIN Politecnico di Torino 001/17
CY CA 01CFIDV 02CFIC OBOTIC RO Burst and Avalanche noise Basilio Bona DAUIN Politecnico di Torino 001/18
CY CA 01CFIDV 02CFIC OBOTIC RO Noise Color Basilio Bona DAUIN Politecnico di Torino 001/19
CY CA 01CFIDV 02CFIC OBOTIC RO Noise Floor (Rumore di fondo) Basilio Bona DAUIN Politecnico di Torino 001/20
Sensors and Mobile Robotics CFIDV 02CFIC CY RO OBOTIC CA 01 Usually the signal noise is modeled d according to a statistic distribution, but Causes of random errors are often unknown or poorly A Gaussian or symmetric distribution is often used, but this can be wrong Example: Ultrasound (sonar) sensors may overestimate the perceived distance, therefore they do not have a symmetrical error distribution Often multiple reflected beams arrive together with direct beams Stereo vision can correlate two images in a wrong way, generating results that are without any sense Basilio Bona DAUIN Politecnico di Torino 001/21